I’d like to bring this experiment back for 2025.

If you’re an early-stage or aspiring deep tech founder and think a 20-30 minute virtual chat would be useful to you, please apply. I’ll hold a total of about 25 sessions over the days before Christmas and New Year’s Eve. If I get a lot of applications, I’ll pick the ones where I’m best able to help.

About me: I was a software engineer for ten years (2nd engineering hire at LinkedIn, then a few years at Google). I then switched to venture capital, co-founded Susa Ventures (software fund) in 2013, and more recently started Susa’s sister fund Humba Ventures (deep tech fund) four years ago.

Criteria:

• You’re US/Canada-based.

• You’re the founder of a pre-Series A deep tech company, or you’re thinking of starting one. (This is the stage where my advice/feedback would be most appropriate.) The areas I’ve been spending time in over the past few years are robotics, energy, defense, manufacturing, space, techbio, etc.

• You have a topic that we could cover well in a relatively short period of time.

Topics that work well: pitch practice/feedback, product feedback, fundraising advice, discussing founder issues, brainstorming product/business/GTM strategy, etc.

Things don’t work well: requests for intros (they don’t scale + it’s hard to give a strong intro after a single short chat), pitching Humba Ventures for an investment (if you want to pitch us, my email is on my LinkedIn page), questions that require deep expertise in a specific area that I don’t know well.

If you’re interested, please apply by 11:59 PM (PST) on December 18. [Sorry, applications are now closed.]

If your friends or online followers might be interested, please share this with them :)

If you love deep tech and are curious about how it looks from a VC’s perspective, then you should check our new fellowship! It’s a program for founders, engineers, scientists, and builders who love frontier technologies and the positive impact they can have. Unlike most VC-sponsored fellowships out there, you don’t need to be interested in a career in VC, and you don’t need to be a future founder that we hope to back.

The fellowship is a 5-week program where you join me and my partner Anna for pitch meetings, internal discussions, and portfolio company catch-ups. We hope to learn from you about the areas where you’re an expert, and you will have fun learning about how VCs work and about mind-blowing technologies like acoustic fire suppression and pet cryopreservation. The anticipated time commitment is 5-10 hours per week.

Here’s a bit about us and the kinds of companies we invest in.

You can find more details about the fellowship and a short application here: https://humbaventures.com/fellowship/

Applications are open until July 19th!

The hardest decisions in life — picking a partner, a job, or a city to live in — all start the same way: with uncertainty and too many options.

While some people stumble into great outcomes by luck, most need a thoughtful process: testing possibilities, gathering feedback, and then committing.

What is explore/exploit?

Explore/exploit is a computer science strategy for balancing two competing objectives:

• Exploration: you want to test new options to gather information.

• Exploitation: you want to commit to the best known option and maximize your expected value.

If you’re building a business, you’ll face an endless stream of explore/exploit scenarios:

• Which market should we target initially?

• Which candidate should we hire as our first engineer?

• Which marketing channel should we focus on?

Because each of these questions is very open-ended, finding the best solution is hard. What’s worse, too little exploration can lead to failure (RIP Quibi), as can too much (e.g. stealth mode startups that never ship). So how do you make a great decision quickly?

How to apply explore/exploit

The explore/exploit approach is to spend some time methodically exploring your options, and then at some point cut off the exploration phase and commit to the best option that you have.

Let’s look at how one might use explore/exploit in three common startup scenarios: choosing a market, hiring, and marketing.

• “Which market should we target initially?” → Make a list of 10 promising markets, cold email until you’ve spoken to at least three people in each market, then focus on the market whose response was the strongest.

• “Which candidate should we hire as our first engineer?” → Solicit resumes until you’ve seen 100 candidates, interview the 10 most promising candidates onsite, extend an offer to the best interviewee.

• “What marketing channel should we focus on?” → Make a list of 10 promising marketing channels, test each of them with a $500 budget, then focus on the channel that performed the best.

Explore/exploit is a powerful framework for navigating decisions with uncertainty, but deciding how long to spend exploring feels more like art than science. Should you try 10 marketing channels, or 3, or 50? Should you spend $500 on each marketing test, or $50, or $50,000? These are important questions, and finding the right balance is especially important for startups, where runway is limited and companies that explore for too long without shipping anything die.

But how do you decide how much time to spend exploring? If you test too few options, you risk making a suboptimal choice. If you test too many, you waste time and lose momentum.

How much should you explore?

If you don’t explore enough, you might be going all in on a bad plan. Perhaps you loved the first engineering candidate and thought they were an A+ so you made an offer, but if you had talked to a few more engineers it would’ve been clear that the candidate was actually a B+ at best.

If you explore too much, you waste time instead of moving forward. If you’ve interviewed 300 engineering candidates and haven’t made an offer, you’re doing something wrong and squandering your time and everyone else’s.

Deciding how much time to invest in exploration depends on several attributes:

• ⏳ Is there a time constraint? If you absolutely need a new engineer within 60 days, then you need to be ready to make offers within a month.

• 🏔️ How high are the stakes for the decision? If you’re betting your company on it, err on the side of more exploration. If it’s relatively minor, go all in sooner and move on to something more important.

• ↕️ Is it more important to capture upside or avoid downside? If you’re hiring a VP of sales, optimizing for the best option matters. If you’re choosing a blog template, good enough is fine.

• 🔄 Is the decision easily reversible? Hard-to-reverse decisions, like hiring or picking a company direction, require more upfront exploration. Easily reversible decisions like buying a monitor do not.

• 🎲 How much variety is there between options? If you’re picking an engineer from a group of 0.5x and 3x and 10x engineers, you’ll need to spend more time exploring than if you were picking from a set of 9x and 10x engineers.

• 🤔 How do you feel about the options so far? If you’ve seen a lot of options and have at least one you’re happy with, move forward. If you don’t like any options so far and you’d regret not exploring more, then explore more.

How to explore efficiently

Set clear boundaries for exploration

1. Establish a clear goal. Without it, exploration is aimless. Do you care more about the cost per user for an acquisition channel, or how many users that channel can reach? If it’s cost per user, is that per user who visits your website, or registers, or makes a purchase? The best marketing channel for maximizing registered users might be very different from the one that lets you acquire purchasing users as cheaply as possible.

2. Estimate the ROI of exploring additional options. Suppose you're evaluating options and assign them scores based on their performance. If the last five scores were 10, 15, 3, 29, and 19 (on some scale), by all means, keep exploring. Maybe the next option will be a 43. But if the last 5 options were 27, 28, 27, 26, and 29, then the variance between options is low, and you should just commit to an option and move on.

Optimize how you explore

1. Balance effort vs. insight for each option. When interviewing job candidates, you don't conduct hour-long interviews with all 100 applicants. Instead, you do 15-minute phone screens with the top 30, then hour-long phone interviews with the top 10, then you bring the top four candidates onsite, and finally you extend an offer to the best one. This incremental approach works in other areas, too. For example, you can spend $200 on each of 10 promising marketing channels, then $2000 on each of the top three channels, and then go all in on the best channel. Frequently, you’ll get enough negative data on an option with the first few minutes/dollars invested, and you can move on quickly.

2. Evaluate options in parallel when possible. For example, if you want to test ten different marketing channels with $500 each, do the tests at the same time instead of one at a time.

3. Prioritize the order of exploration. If you have a lot of options, start with the most promising ones first. For example, if you are considering 20 marketing channels and test four at random, their CACs (customer acquisition costs) might be $20, $30, $7, and $11. This is a wide range, and it suggests that more testing is needed. But if you rank the channels by their perceived potential and test the most promising four, perhaps their CACs will be $2, $4, $9, and $7. Now you might feel comfortable stopping the exploration phase and picking the channel with the $2 CAC.

Explore broadly and learn from others

1. Sample from different types of options. If you’re interviewing engineers, don’t restrict yourself to your classmates from Harvard class of ‘08. Instead, cast a wide net and consider past colleagues, cold inbound candidates, cold outbound candidates, candidates from sites like interviewing.io, bootcamp graduates, and so on. Similarly, if you’re trying to pick a marketing channel, test the channels you’re the most comfortable with, but also channels used by your competitors, channels ignored by your competitors, and channels that your audience is likely to see. The more starting points you explore, the less likely you are to climb the wrong hill and get stuck in a local maximum. The outcome of the exploration phase is only as good as the quantity and diversity of the inputs.

2. Learn from others. If you’ve never hired a salesperson before, you can spend 100 hours interviewing candidates and slowly calibrating your understanding of sales ability. Or you can ask a few of the best salespeople you know (and perhaps the best salespeople that they know) to share which attributes they consider to be important.

“Only a fool learns from his own mistakes. The wise man learns from the mistakes of others.”

- Otto von Bismarck

Common traps

The most common mistakes are not exploring enough or exploring too much.

• If you’re avoiding exploration, ask yourself if you’re picking the best option or just the easiest one.

• If you’re stuck in endless testing, consider whether new options are meaningfully better, or if you’re just afraid to commit.

Not exploring enough usually stems from avoiding discomfort. People like easy choices, and so picking the option that seems good enough or the one that you’re most comfortable with is tempting. If your decision is important, ask yourself if you’re choosing the best option or if you’re taking the easy way out.

Exploring too much usually stems from indecisiveness. If every new option you try is 10% better than the last, great, keep exploring. But if each incremental option provides very little value then perhaps you’re anxious about committing to a single choice. Ask yourself if you’re allowing the perfect to be the enemy of the good.

Putting it all together

As a founder, you’re paid to make decisions under uncertainty. Mastering the explore/exploit tradeoff isn’t just a clever mental model — it’s a survival skill. Know what you're optimizing for, follow the recipe below to explore intelligently, and then commit with confidence.

Here’s a concise guide on navigating this balance:

1. Decide on the metrics you’re optimizing for.

2. Allocate some amount of time to explore your options. If a decision is more important, less reversible, or involves wildly different options, commit to spending more time exploring.

3. Prioritize options and explore — in parallel if possible — until time runs out or you’re happy with the best option.

4. Double down on the best option.

This insight has driven Humba to partner with founders building everything from organoid-based drug testing platforms to next-generation sawmills to nuclear power for defense. Many companies seeded by Humba’s first fund, like Antares, WindBorne, and Mytra, have already gone on to raise subsequent rounds from top firms like Founders Fund, Khosla Ventures, and Coatue.

We recognize that real growth is driven by innovators building at the fundamental substrate of our economy — the physical world. That’s why our fund strategy invests in domains like transportation, energy generation and distribution, heavy industry, health care, and national security. Innovators building in these spaces often carry more technical risk than market risk, and we embrace that. The question for us is not whether there will be demand for cancer cures or cheaper energy, but what it will take for teams to develop them.

Luckily, we believe it’s never been easier for a small group of builders to make a massive impact. Ideas that used to require tens or hundreds of millions can now be approached with a few million dollars or less. Humba supports founders at the earliest stages of this journey. With Humba II, we are investing $250K-$500K at pre-seed and $500K-$1M at seed. And, because we’re part of the broader Susa fund family, we offer more platform services and support per dollar invested than much larger funds.

Companies we partner with often require teams with diverse, unique backgrounds, and we are building our team similarly. Humba’s General Partner, Leo Polovets, was one of the first engineers at LinkedIn and has over a decade of experience in venture, having co-founded Susa Ventures and helped build it into the leading seed fund it is today. Humba’s newest addition, Anna-Sofia Lesiv, has excelled at distilling deep tech topics to mass audiences through her research and writing.

History has shown us that if you just get enough talented engineers in the same room together and give them enough capital and support, magic will happen. By definition, building hard tech is hard, but nothing that endures comes easy.

For founders eager to take bold swings, Humba is ready to take on the risk with you. If you have a big idea, we want to meet you. If we invest, we’ll commit to supporting you throughout your entire journey.

This post was originally published on Humba’s blog.

The biggest overall misconception about deep tech companies is that they are necessarily much harder and more expensive to build. That may have been very true 10-20 years ago, but it’s less true now. This post will make the case that building these companies is easier and more affordable than ever before. (Note: still not easy, but getting easier!)

There are several reasons why this is a great time to build a deep tech company:

1. Stronger startup talent

2. Improved tech infrastructure

3. Better access to venture capital and non-dilutive funding

4. Strong macro tailwinds

Startup talent

Twenty years ago, companies in defense tech or advanced manufacturing or tech bio had a much harder time with recruiting. Back then, it was hard to find candidates with domain experience and a startup mindset. For example you could either find a startup engineer who didn’t understand how the defense sector worked, or an engineer from a giant defense prime who didn’t know how startups worked. It was very hard to find someone with a foot in each world.

Today, incredible alumni from companies like SpaceX, Palantir, Solugen, Anduril, and so on are available as co-founders, advisors, and employees. It’s no longer impossible to recruit a team of ambitious engineers who can build complex hardware+software products at startup speeds.

Additionally, by hiring startup people with extremely relevant industry experience, you can learn from recent playbooks that have worked instead of reinventing every wheel yourself.

Tech Infrastructure

Building a deep tech company today is easier because the “primitives” in most areas have gotten dramatically faster/smaller/cheaper over the past few decades. We talk about Moore’s law for computing, but there are Moore’s-law-like trends for many fundamental hardware and bio primitives.

The trend looks a lot like this:

Examples of technologies that have advanced a lot relative to where they were 15-20 years ago:

• Solar panels are 10x cheaper per watt1 and 1.5x-2x more efficient.2

• Lithium-ion batteries are 10x cheaper3 and 10x more energy dense.4

• LiDAR is about 100x cheaper.5

• GPUs are 1000x cheaper per GFLOP6 and 2000x more powerful.7

• Genome sequencing is 100,000x cheaper (!!) and gene synthesis is 100x cheaper.8

• Launching a kilogram into space is 10x cheaper.9

• Industrial robots are 100x more precise than 40 years ago, 5x cheaper than 20 years ago.10

• Hard drives and RAM are each 40x cheaper and much higher capacity.11 And there are amazing options if you’re space constrained (a few of these in 2004 vs one of these in 2024).

Across the board, key components for hardware/bio/etc are frequently 10x+ cheaper, 10x+ smaller, and 10x+ faster today than in the early 2000s.

Additionally, “software for hardware” is getting better as well. Creation of synthetic data for testing is better, collaboration tools are better, physics tools are better, design and simulation tools are better, and so on.

What this means is that a typical deep tech company might have needed tens of millions of dollars to prove out something meaningful 20 years ago, single digit millions 10 years ago, and only hundreds of thousands today.

For example, I recently spoke with a company that enables biology experiments in space. They’re less than a year old and have already launched multiple experiments into orbit with a small pre-seed round (!). That would’ve been impossible in 2010 without lots of funding — and maybe impossible even with lots of funding. In general, many amazing deep tech startups today couldn’t have been built until the last few years, but now founders can build compelling proof-of-concept hardware with a pre-seed or seed round.

Venture Capital and Non-dilutive Funding

In addition to better tech components and stronger startup talent, there’s currently a lot of funding available for deep tech.

On the VC side, you have large funds like DCVC ($4b AUM), Lux ($5b AUM), and Eclipse ($4b AUM) that by themselves have invested billions into deep tech companies over the past decade. You also have great seed stage funds like Root Ventures, Fifty Years, and Cantos, and emerging funds like Humba Ventures (my fund!), Champion Hill, Julian Capital, Also Capital, and Undeterred Capital. And there are hundreds of other great funds that hopefully won’t be mad at me for keeping this list short, as well as deep tech demo days like Boost and pitch forums like Deep Checks. Almost none of these funds and programs existed in 2010, and many didn’t exist until after 2020. Thanks to recent mega-successes like SpaceX and Palantir, the amount of well-deserved investor attention that deep tech is getting right now is insane.

Of course, selling a piece of your company for every dollar you have to raise is not ideal, and that’s where non-dilutive funding comes in. The last decade has given rise to a lot of government programs like SBIR grants, AFWERX, STRATFI/TACFI matching funding, and the DIU. Additionally, startups in areas like climate and energy can benefit from generous incentives in the Inflation Reduction Act, and from DoE grants/loans.

The cash that these government programs offer is great and helps keep dilution at bay. But the programs are also wonderful because you can get early access to your customer if you sell to government departments. Until recently, you had to make it on your own for years before the DoD would talk to you, but now they roll out the red carpet and engage with companies and founders within months of inception. This type of early customer development opportunity is invaluable.

Macro Tailwinds

Talent and funding are great, but they won’t get you anywhere if customers don’t want to buy what you’re selling. And that’s where recent economic and political tailwinds come in. Tensions with China and Russia have created pressure on our government to fund defense innovation. Strained relations with China are also causing companies to rethink their material sourcing and manufacturing for the first time in years or decades. Additionally, rising minimum wage laws and intensely seasonal hiring needs in sectors like agriculture are pushing many industries toward increasing automation in order to survive.

These global trends are an immense tailwind for startups focused on robotics, manufacturing, defense, energy, and many other areas.

There are now lots of buyers that had zero desire to try anything new 5-10 years ago, and that are now desperately searching for alternatives to their status quo. That’s an amazing opportunity for companies that listen to their customers and build quickly.

It’s time to build deep tech!

The confluence of dropping costs and improving performance of tech components, tons of talented deep tech startup alumni, lots of early stage capital, and strong global tailwinds have created an excellent environment for building deep tech companies.

Software will continue to be a great place to build and invest, but I believe the physical world will present even better near-term opportunities for ambitious founders. Ideas that used to require 50 people, $50m, and 5 years can now be tested with 5 people, $2m, and 2 years — and we’re going to see a lot more amazing technology in the physical world as a result.

Footnotes

I’ve now been a VC for almost eleven years. While most of my investing career was focused on B2B SaaS investments at Susa Ventures, two years ago I started pivoting into deep tech investing1 via our new fund, Humba Ventures. As I started investing in areas like robotics, energy, biotech, and defense, I tried to learn more about their historical performance.

Since I’m an engineer and data guy by training (and since I’m betting my career on these categories!) I decided to dig into some PitchBook data2. This exercise convinced me that deep tech is the best place to invest and build right now, and that the following four assumptions about deep tech companies turn out to be misconceptions:

1. Deep tech companies have poor outcomes.

2. Deep tech companies are much more capital intensive.

3. Deep tech companies take much longer to exit.

4. Deep tech companies have much higher failure rates.

The rest of this post covers each of these points in detail.

Do deep tech companies have good outcomes?

Yes!

Many deep tech industries are huge, and there’s a lot of room for new players: batteries ($100b market), pharma ($1.5T), defense R&D ($150b in the US), industrial and warehouse automation ($200b total), and so on.

To take defense as a concrete example, this site lists awarded defense contracts. The DoD awards multiple 8-9 figure contracts daily. Any startup that captures just a few contracts of this size is basically a $1b+ company.

To get more specific, here’s a sample of great exits from just the past five years:

- 10x Genomics ($4.7b, gene sequencing tech)

- Archer Aviation ($1.7b, electric aircraft)

- Auris ($3.4b, surgical robots)

- Cylance ($1.4b, antivirus)

- Foundation Medicine ($5b, genomic profiling)

- IONQ ($2.3b, quantum computing)

- Luminar ($1.5b, LIDAR)

- Mazor Robotics ($1.6b, surgical robotics)

- Nextracker ($9b, solar panel positioning)

- Nuvia ($1.4b, chipmaker)

- Planet Labs ($2.8b IPO, satellite imagery)

- Quantumscape ($3b, lithium battery developer)

- SentinelOne ($5b, security platform)

This doesn’t include many valuable still-private companies like Anduril ($10b), Locus Robotics ($2b), Redwood Materials ($5b), Relativity Space ($4b), Solugen ($2b), and, of course, SpaceX ($150b).

I think we’ll continue to see more and bigger outcomes in these spaces in the next few years. Many industries and problems could benefit from deep tech innovation, and either have no incumbents or old incumbents. For example, ABB is a leader in industrial automation ($30b annual revenue). It’s also a merger of two companies founded in the 19th century. It feels unlikely that in the next decade, the best entrepreneurial minds from SpaceX, OpenAI Microsoft, Tesla, etc., will be unable to make inroads against ABB’s market leadership.

Aren’t deep tech companies much more capital intensive?

On average, no! Lists of “most funded companies” are almost exclusively more traditional startups in areas like marketplaces, fintech, social networking, etc. These are companies like Uber, WeWork, and Oscar, which spend orders of magnitude more than most deep tech businesses.

Here’s a summary of PitchBook data for US/Canadian companies founded since 2010 that have had $250m+ exits3:

My analysis found that companies across traditional and deep tech sectors are similarly capital intensive. The median company has raised 11-15% of its exit size in prior funding, which runs counter to the common assumption that to get a $1b exit in deep tech, you need $500m+ in funding. Reasons for deep tech companies being less capital intensive than expected are an increasing abundance of grants and other non-dilutive funding, less competition for new customers, and larger contract sizes.

The table above also reinforces that there are many meaningful exits across all of these categories. They are all good territories for founders and investors to explore!

Don’t deep tech companies take much longer to exit?

The table above shows that the deep tech and life science companies actually exit 10-25% faster than their traditional counterparts. For example, the average company age for exits above $500m is just under 8 years for traditional companies, just over 7 years for deep tech companies, and under 6 years for life science companies.

My hypothesis is that deep tech companies that build valuable IP can have meaningful exits quickly, while traditional tech companies rarely have very strong IP and are mostly acquired based on great traction. It often takes longer to get to meaningful traction than to accumulate valuable IP.

(On a side note, this is why deep tech VCs value strong founding teams even more highly than traditional VCs: if the team is good enough to produce unique, valuable IP, then it’s likely to have a good exit.)

Isn’t the failure rate especially high?

For me, this was the most interesting area to dive into. I got data across about 15 sectors for companies started since 2010 and with at least $1m in VC funding. My goal was to estimate the chance of a company with at least some VC funding eventually reaching either a $250m+ valuation or a $250m+ exit. I picked $250m as the threshold because that is already a good outcome, and at higher levels like $1b the data gets too sparse.

Here’s a table of industries, and the relative frequency of $250m+ valuations given at least some VC funding:

This table is sorted by the odds of hitting $250m+, and you can see that most categories are in the 7-9% range, with a few outliers on both sides. Deep tech sectors fare quite well here (except for AR/VR 😢). Basically, almost regardless of category, your chance of hitting a $250m+ valuation is around 7-9%.

Of course valuations are secondary to exits, so here’s a table of relative frequencies of $250m+ exits in each category.

Here deep tech actually dominates traditional tech. Companies in areas like defense, space tech, and life sciences have 2%-5% odds of having $250m+ exits, while companies in categories like SaaS, Fintech, and AI/ML have 1%-1.5% odds.

If you look carefully at the table above, you might notice that the odds of good exits in deep tech categories are often higher, but the number of good exits in these categories is lower. I think that’s a hint about why these companies tend to be more successful. Building a deep tech company is hard. Really hard. You typically need a multidisciplinary team that’s exceptional in multiple areas, and that’s a very high barrier to entry. But if you can assemble such a team then you will have much less competition than most startups and a higher chance of a good outcome.

For example, we recently invested in Antares, which is developing nuclear microreactors for defense applications on earth and in space. A good team for a company like this needs to know how to build reactors, how to sell to the defense department, how to deal with regulations, and so on. It’s very hard to assemble a team like this, and that’s why there are only a few microreactor startups. If you’re good enough to enter the space and get funded, you’ll be one of 1 or 3 or 10 companies fighting for market dominance, not one out of 500. This leads to better expected outcomes for the founding teams that are strong enough to get funded in the first place.

Key Takeaways

To summarize the data above:

• There are lots of $1b+ deep tech exits.

• While the absolute number of strong exits is smaller than for traditional sectors like B2B SaaS, the rate of good exits is approximately 2x higher.

• Good exits for deep tech companies, on average, come 6-24 months sooner than good exits for traditional tech companies.

• Deep tech companies, on average, have similar capital intensity to traditional tech companies. Across all exits above $250m, the average amount of capital raised is about 11-15% of the sale price regardless of sector.

Finally, while this post focuses on the financial performance of deep tech companies, investors and founders should also consider the opportunity to have a monumental impact on the world. Most startups are highly impactful on some group of people, and for deep tech companies both the impact and the size of the group tend to be much larger. Now is a better time than ever to build deep tech companies: off the shelf technology keeps improving, there are lots of funding and grant options, there are many successful deep tech companies to poach talent from, and macro trends like reshoring provide strong tailwinds for ambitious ideas.

If you’re building or investing in these spaces, please reach out! I’m leo@humbaventures.com, and our fund Humba Ventures is investing $250k-$500k in areas like energy, robotics, material science, and defense and aerospace.

Thanks to Julian Shapiro, Heston Berkman, Mike Annunziata, Mr. Exits, Jai Malik, Ian Rountree, Chelsea Goddard, and Josh Manchester for feedback on this post.

Sometimes companies run out of money because they never had enough capital for their plan in their first place. But sometimes bankruptcy is self-imposed by careless spending, weak decision-making, or poor financial hygiene.

This post is a collection of cash management heuristics, tips, and strategies that I’ve accumulated over the past decade of investing in and working with 200+ startups like Robinhood, Flexport, and Mux. While these tips are aimed at venture-backed startups, many of them apply to bootstrapped startups as well.

Core Framework

In order to optimize your cash position and your odds of success, you should focus on four key pillars, which I call the ARGO framework:

💰 Acquire more cash. A great way to avoid running out of money is to get more through various forms of funding.

✂️ Reduce costs and push them back. The lower your costs and the later you can pay them, the better that is for your cash position.

🌱 Grow revenue and pull it forward. The more you sell and the faster the money lands in your bank account, the better that is for your cash position.

⏱️ Optimize operations. A great way to not run out of cash is to get more leverage out of each dollar through improved efficiency.

The ARGO framework helps improve your odds of surviving and thriving from a cash standpoint.

Now, let’s dive into tips and mental models for each of these areas.

💰 Acquire More Cash

• When fundraising, add 6 months of cushion to your fundraising target. In the worst case, you will need 6 months to either raise your next round or run an M&A process. In the best case case, you will have more runway to hit break-even. You don’t want to hit all of your target milestones the day you run out of money because, well, you’ll be out of money.

• If you’re currently fundraising, raise the right amount. If you need $3m to execute on your plan, don’t raise $2.5m — that’s a lot of dilution to take and it won’t be enough to hit your target milestones.
  
  A good financial model can help you triangulate the proper round size. Some tips for building a good model can be found below:

• The worse the fundraising climate, the more willing you should be to raise money on any reasonable valuation. The better the climate, the more picky and reluctant you can be.
  
  Concerns about dilution should be inversely proportional to your likelihood of running out of money.

• Take advantage of startup tax credits with services like MainStreet or Neo.tax. Tax credits are free money, and they’re surprisingly generous. Both Canada and the US offer R&D tax credits — and I’m guessing other countries do, too.

• Put your money into a safe, high yield account like Vesto.1 You can earn 4.5+% on your idle cash using highly liquid, low risk investments like US treasuries and CDs. If you have 24 months of runway and you put your cash into a high yield account, you'll get an extra month of runway for free.
  
  Avoid risky investments like high yield crypto schemes unless you want to gamble with a small portion of your cash.

• Send regular investor updates. Investor updates are basically a customer success tool. If you send updates then there’s a much higher chance that your existing investors will invest in your next round; if you never communicate, existing investors become a poor source of future funding.

• Use cards like Ramp or Brex to earn rewards on your corporate spending. It’s an easy way to get 1%-2% back on your costs.

• Consider venture debt or non-dilutive revenue-based financing. Venture debt is usually tied to a VC funding round. Revenue-based financing becomes available around $50k+ MRR, and you can typically borrow 4-8 months of MRR at a 10%-20% interest rate. That’s not cheap, but it’s not bad if your company is growing rapidly.2

• If you’re about to fundraise, work on your storytelling skills, pay $1k-$2k for a designer to polish up your deck, and practice your pitch a lot. The ROI on improving your fundraising skills is immense given that an investment of a few dozens hours can lead to hundreds of thousands or millions more dollars for your company.

✂️ Reduce Costs

• Make cost-cutting decisions quickly. The earlier you cut a cost, the more money you’ll save.

• A lot of companies increase burn quickly after raising around. One joke in the VC industry is that no matter how much a startup raises, the money somehow always lasts about 18-24 months. Stay lean before you have product/market fit and resist the temptation to spend quickly.

  A good rule of thumb is to set a cap on net burn, like $100k/mo. If you want to spend more than $100k/mo, you’ll have to make up the difference with revenue. One perk of this approach is that runway calculations become easy: if you raised $3m and you are capping net burn at $100k/mo, then you have at least 30 months of runway.

• Unfortunately the most effective way to save a lot of money is often to do a layoff. If you have to do layoffs, make sure you only do them once. Three rounds of 10% are much worse than one round of 30%. Morale gets crushed with multiple rounds of layoffs, and you save less money because the rounds happen over months instead of all at once.

• When possible, prioritize contractors and on-demand help over full-time employees. Do you need several full time designers or just one? Do you need a full-time designer, or would a design contractor suffice? Do you need a contractor, or would an advisor be enough? Finally: do you need an advisor, or is the design good enough for the foreseeable future?

• Be frugal. Negotiate what you pay to your vendors, leverage discounts that you get from investors or founder communities, and so on. For example, at Susa Ventures we have an internal Notion doc for with about 100 discounts we’ve negotiated for cloud infrastructure, HR tools, sales tools, corporate insurance, you name it. If you’re going to spend $250k this year on products like these, why not look for discounts and save $50k-$100k?

• Think about “dead time” when hiring employees. Don’t hire a front end engineer 3 months before you’re ready to build the front end. Don’t hire a salesperson if they’re going to be ready to sell right when your business has a seasonal slowdown. Time your hires so that you can maximize the value that each person brings to the company.

• Prefer renting/on-demand over buying, especially when you have limited runway. Let’s say you need a server that will last for 5 years, and you can buy it for $5k or rent it on AWS for $150/mo. If you have 12 months of runway, rent! If you have 3+ years of runway then buying might make sense. Maybe.
  
  TCO (total cost of ownership) calculations are useless if the lifetime of owning something is much longer than your company’s runway.

• Run minimum-viable experiments. Do you need to spend $20k to see if AdWords is a good channel, or can you start with $2k or even $200? If you’re not sure if you need an EA, can you start with a temp or Athena instead of hiring someone local?

• Defer costs as much as possible. It’s better to pay $1k/mo for 12 months than to pay $12k upfront. It’s even better to pay $12k at the end of twelve months — even if there’s a small additional penalty for that. The later you pay for things, the more cash you have remaining today. Services like Ramp Flex can help here.

🌱 Grow Revenue

• Prioritize customers that will pay you more, or that don’t mind paying you upfront. Revenue is the opposite of costs in terms of timing: $12k in revenue today is better than $1k per month, which in turn is better than $12k at the end of the year.

• Prioritize customers with a shorter payback period. A $200k ACV customer with an 18-month payback period is worse than a $50k customer with a 1-month payback period.

• Alternatively, give customers a discount for paying up front, or use a factoring service like Pipe. A factoring service will give you ~90 cents on the dollar for the value of a contract, and then they’ll collect the future revenue on that contract.

• Look for easy wins to expand your customer base or increase your ACVs. For example, can you sell your product internationally by outsourcing a lot of the complexity to Stripe? Can you create a subscription version of your product if it’s normally a one-time sale?

• If you’re in a real bind, consider nontraditional or unscalable revenue sources. Anything from consulting work to selling cereal. Just remember that you need to make enough money to eventually invest in the core product, otherwise you’re just treading water and postponing the inevitable.

• Experiment with pricing! Most companies undercharge because they take a guess at pricing at launch and then that’s their price forever.
  
  Generally speaking, you should try to capture 15-25% of the value you create. E.g. if you can help a customer generate $100k in extra profit, you should be charging ~$20k (not $2k and not $80k). A few pricing resources: Monetizing Innovation and Pricing your SaaS product.
  
  One good pricing heuristic: keep doubling your prices after every few sales until the majority of prospects push back.
  
  This is probably my favorite tip because a startup can 2x or 5x its revenue in a short period of time just by being smarter about how its products are priced.

⏱️ Optimize operations

• Experiment, experiment, experiment. Experiments will help you grow and optimize your business. For example, spending $25k on an AdWords campaign probably performs worse than spending $5k to try ten campaigns and then spending an additional $20k on the best one.

• Parallelize work as much as possible. For example, do you have to wait for your engineering team to finish building a feature to start creating marketing copy for it? Or can you create copy at the same time as the feature is being developed? If you do tasks serially instead of in parallel, then you’ll have teams of people waiting on the output of other teams of people instead of actually doing work.

• Use funnel optimization tools to improve your sales and marketing conversions. Basic off-the-shelf tools can quickly improve your sales by helping you identify basic issues in your existing GTM playbook.

• Maximize your “at bats”. Let’s say your targeting becomes 1% better every time you publish a newsletter post and track the results. If you write a monthly newsletter, your targeting will improve by about 12% this year. If you publish weekly, your targeting will improve by almost 70%. The more “launch → get feedback → iterate” loops you can execute, the more progress you’ll make.

• If you’re dependent on future fundraising, optimize for “ROI before the next fundraise.” For example, if you have 6 months before you need to raise, don’t point your entire sales team on prospects that will take a year to close. Similarly, don’t hire someone who will need at least X months to ramp up if you have to raise in <X months — that just increases your burn but won’t get your company to a better spot before your raise.3

• Identify and address your largest bottlenecks. If you get 5 leads per month and 80% of them turn into customers, then it’s much better to try to grow 5 leads to 50 than to try to improve your close rate from 80% to 100%. But if you have 5,000 leads per month and 1% of them turn into customers, then you should focus on improving your close rate (or figuring out why your leads are so poor), not on growing the number of leads. Find your bottlenecks and eliminate them.

Parting thoughts

To recap the ARGO framework:

💰 Acquire more cash through fundraising, revenue-based financing, factoring, and high yield corporate accounts.

✂️ Reduce your costs by renting instead of buying, cutting costs quickly, and running small tests before going all in on expensive initiatives.

🌱 Grow revenue by expanding who you sell to, increasing prices, and prioritizing customers with favorable economics.

⏱️ Optimize operations by doing lots of experiments, using off-the-shelf tools to improve your sales and marketing funnel, and moving quickly.

Finally, while the advice above will help you manage your startup’s money better, it’s important to remember that “you can't cut your way to success.” You can extend your runway and be more efficient, but ultimately you have to create something that customers want and will pay for. If stretching your capital further helps with that, then great. But if you feel like extending your runway is just postponing an inevitable shutdown, then stop trying to extend your runway.

Most of us get 40-50 years in a career, if we’re lucky, and every year you spend on something that’s not going to work out has a huge opportunity cost. So be frugal, experiment, and optimize your cash flows, but don’t forget that the ultimate goal is to build a high impact, lasting business. If you’re not progressing toward that goal then it’s okay to change direction.

If you enjoyed this post, I recently gave a talk hosted by Stripe on how startups can weather the current economic climate. The recording can be found here.

Being a VC involves looking at thousands of pitch decks annually. It’s a fun part of the job, but because I rarely have a founder’s voice-over while reading a deck, my interpretation of any given slide can be very different from what the founder intended.

I was helping a friend iterate on their pitch deck last week. One of their slides had a table that showed the evolution of a handful of metrics over time:

Before you keep reading, spend a few seconds looking at the data. How would you summarize this company’s performance?

Here is the dialogue that ensued between me and the founder:

Me: Is this the best way to present the data?

Founder: What do you mean?

Me: Well, it’s nice that your churn rate is dropping and revenue per user is going up, but the other data makes it look like the company is doing poorly.

Founder: It does? How??

Me: First of all, the number of new users peaked a few months ago, so it looks like growth is decelerating. Is that the case?

Founder: No. That number fell because we removed our free plan in the middle of June, so now all new users are paid users. Before, we had a lot of people who signed up on the free plan and never got serious about the product. So we get fewer new accounts now, but they are much higher quality.

Me: Ok, that makes sense. How about the customer acquisition cost? It was dropping a few months ago, but now it’s rising sharply?

Founder: That’s related to us removing the free plan. Because the value of the average user is now much higher, it made sense to have a higher budget for acquiring new users.

Me: Ah, I see. I guess I was mistaken about revenue and churn trends being weak. What about the amount of time users spend using the product? It can’t be good that your customers are using the product less and less every month?

Founder: It’s great, actually. It took us a lot of work to get to this outcome. We realized that our website was slow and that was contributing to user unhappiness. We spent the last 6 weeks making pages load 10x faster, and now users are much more engaged. Obviously that made session times drop, since users spend much less time waiting for things to load.

At this point in the conversation, I pointed out that all of the founder’s explanations were reasonable and good, and the business is doing great, but no one who read the deck by themselves would know these backstories. If data and graphics are presented without commentary, some readers will have a generous interpretation while others will have an uncharitable one.

If you want your reader to draw a conclusion, just tell them the conclusion! Don’t beat around the bush. If your data shows that growth is decelerating, or churn is spiking, or your team is shrinking, then your audience will assume the worst in the absence of more information. Given them the information they need to contextualize your data.

If I were to rewrite the founder’s metrics slide above, I’d just turn it into a few explicit bullet points:

Strong progress since April:

• Dropped free tier, which doubled our per-user revenue.

• Sped up our site 10x, which helped cut churn from 10% to 5%.

• Revenue/Acquisition Cost rose from 2.5x to 3x. We anticipate 4.5x by EOY.

This version of the slide clearly states the things that the founder is trying to get across, and it doesn’t present extraneous information that can be misinterpreted.

Simple tips for removing ambiguity

Whether you’re writing an article or creating a pitch deck or crafting your resume, you should leave as little as possible to chance. If you suspect there’s an ambiguity, eliminate it.

• The title of each slide in your deck should be the key takeaway. Replace “Traction” with “6x growth in the last 8 months.” Instead of “Our Team,” use “Rare mix of Oil&Gas + Marketplace experience.” This is a good practice for deck, memos, blog posts, etc.

• For each slide, ask yourself what the worst possible interpretation would be. If your competitor had to make a damning presentation based on your materials, what would they say? How would you change your deck to ruin their attempted sabotage?

• Ask 5+ people to write down their takeaways and gut reactions to each of your slides. Identify anything that seems unclear and fix it. If some people think you’re trying to say X while others think you’re trying to say Y, then the slide is broken.

Types of bottlenecks

There are three common ways in which startup leaders can fail to scale up:

1. They’re good at some parts of the job but bad at other parts. Jobs often require multiple skills, and key employees have to be great at all of those skills. For example, an early-stage CTO should be a great engineer, a great recruiter, and a great manager. If they’re not good at one or more of these things, the engineering team will struggle as it grows.

2. They’re unwilling to delegate. In growing companies, strong performers gain responsibility quickly. If you’re good at your job you can quickly move from individual contributor to manager to executive. This transition requires letting go of making personal contributions, and people are often reluctant to do that. A company ends up having a CEO who wants to spend 10% of their time defining the product strategy, or a CTO of who manages 8 people but still wants to check in code for core features. The result is that these people become bottlenecks. The CTO becomes rusty and their code quality worsens, but they also hold up their teammates who are waiting for work that the CTO committed to but doesn’t have time for.

3. They’re good but not great at the job. Some people simply hit a limit in what they’re able to do. Maybe someone is good at managing 3 people but not 8 people. Or they work well with managers, but not managers of managers. If these people don’t improve their skills, they end up holding the company back. “Good” may be good enough on a small team, but it’s not sufficient when dozens or hundreds of people depend on a single person. The danger to a company is even greater if a key person isn’t great at their job but thinks they are.

Are you a bottleneck?

One of the challenges with scaling up is that people often don’t realize that they’ve hit a wall. They get promoted from contributor to manager to director and think everything is going great, when in fact they’re no longer performing well at their new level. It’s hard to stay self-aware when your environment is changing rapidly. Here are some tips for figuring out if you’re transitioning from an asset to a bottleneck:

• Listen to your gut. If you feel like you’re not performing adequately, don’t just ignore that feeling. Look for data that tells you if you’re just feeling impostor syndrome or if you’re actually underperforming.

• If you trust your colleagues to be honest and direct with you, do 360° reviews. Ask your peers, reports, and managers about what they perceive to be your strengths and your weaknesses, as well as advice on how you can improve.

• If you think people are afraid to speak up, solicit anonymous feedback through a survey or a comment box.

• If you’re a key exec at your company (CEO, CTO, VP of Sales, etc), consider doing sessions with an executive coach. Coaches can be an incredible asset for leveling up, and a good coach will provide you with impartial feedback since they have no stake in your company.

• If you’re new to a job type or job level, talk to 5-10 peers that are great at your job. Understanding how other people perform and how you stack up against them will help you assess your strengths and weaknesses.

A thought experiment on personal scaling

The best way to avoid being a bottleneck is to not become one in the first place. One thought experiment that I sometimes suggest to founders is to consider what their future selves will have to do.

Thought experiment: What will your job look like in 1 year? In 5 years? Which existing responsibilities will you no longer have and what new responsibilities will you accumulate?

This perspective can make your career path clearer. Here are three examples of how responsibilities may evolve over time:

1. You’re an early-stage CTO who currently splits time between managing 3 engineers and writing code. What will your role look like in a year? Your team will probably grow to 5-10 people, so you’ll have more management responsibilities and less time to code (if any). What will your role look like in 5 years? If the company does well, it might have 100 engineers. At that point, you’re either purely a manager or purely an engineer, and very unlikely to be doing a mix of both.

2. You’re a CEO and you are starting to build out a sales team. Your company is young, so you’re doing most of the selling. You’re also personally training each sales hire. What will your CEO role look like in a year? As your sales team grows, you’re probably doing less selling and training, and focusing more on hiring a VP of Sales. In 5 years, you’re likely delegating everything sales-related to your VP of Sales.

3. You’re an engineer at a Series A company, and your team is doubling every 6-12 months. What will your role look like next year or in 5 years? Your department might be 2x-4x bigger next year and 20x or 50x bigger in 5 years. You have a lot of career options in a company that’s growing that quickly: remain an individual contributor, become a tech lead, shift from engineering to product management, become a manager, or even work your way up to VP-level. Each of these options requires a different, intentional plan of action, and if you’re not pursuing a specific plan then there’s a good chance you’ll still be an individual contributor in 5 years. Maybe that’s what you want, and maybe it’s not.

Becoming your future self

Once you have a sense of how your role is likely to look in 1-5 years, the next step is reverse engineering a path that gets you to that destination. There are 4 main courses of action:

• Improve. As you go from managing two people to managing ten, or from recruiting managers to recruiting directors and VPs, you’ll have to level up your skills. You can do this by talking to peers who are several years ahead of you, reading books, going to conferences, and becoming more self-aware of your strengths and weaknesses.

• Transform. If your role will shift over time (e.g. from individual contributor to manager, or from selling to recruiting), you will have to transform your skill set accordingly. Talking to peers and studying for your next role in advance is a great way to be prepared when you make the change.

• Delegate. As the number of people reporting to you grows, certain tasks will fall off your plate. Are you giving every new engineer a 4-hour walkthrough of your proprietary mobile app framework when you have 3 employees? Probably. How about when you have 30 employees? Maybe. How about 300 employees? Absolutely not. As you consider the tasks that Future You will not be doing personally, start delegating those tasks to the appropriate people. The sooner you delegate, the more you can focus on the things that Future You should prioritize.

• Step aside. Sometimes thinking about your future at a company will help you realize that you don’t like that future. In those cases, the best course of action might be to step aside for someone who is a better fit. This might mean changing your role at a company, like shifting from CEO to VP of Product, or from tech lead back to individual contributor; or it could mean going to another company where you can do the work you want to do. Leaving a company is a last resort nuclear option if you’re a founder, but very viable if you’re not.

Miscellaneous tips

As your role evolves, one of the toughest things to do is to prioritize the right tasks and activities. Here are a few assorted tips for doing that:

• The more you can delegate or automate your existing tasks, the more time you’ll have for focusing on upcoming projects.

• Talk with peers you respect and get advice from them about areas where they have the experience that you lack.

• Consciously think about where your role is going on a regular basis. Ideally every 2-6 months if you’re at a fast-growing company.

Perhaps the most important thing to consider is that if you don’t actively think about what Future You will be doing and how to get from where you are to where you’ll need to be, then Future You will look a lot like Present You.† If you want to scale in tandem with your company, then start moving in the right direction today.

† This is true professionally and personally.

Thank you to Natalie Dillon for giving me feedback on this post.

Table of Contents

1. Audience Growth: The Little Things that Work by Sam Parr, Founder of The Hustle

2. How to Scale a Content Team by Maggie Leung, VP of Content at NerdWallet

3. BuzzFeed: How We Data Science by Gilad Lotan, VP of Data Science at BuzzFeed

4. Why Data-Driven Companies Hire Economists by Jed Kolko, Chief Economist at Indeed

5. Attract, Capture + Convert Customers by Shafqat Islam, CEO of NewsCred

6. The Anatomy of a Viral Video by Grant Marek, Editorial Director at Chubbies Shorts

7. The POPSUGAR Formula by Lisa Sugar, Founder of POPSUGAR

8. Fireside Chat: Marketing for Startups with Michael Seibel, CEO of Y Combinator

9. From Zero to Five Million Pageviews by Courtland Allen, Founder of Indie Hackers

10. Connecting with Journalists by Ricky Yean, CEO of Upbeat

11. Hard-to-Make Data-Driven Content by Amber Thomas and Matt Daniels of Polygraph

12. Content that Caters to Multiple Tiers of Media by Issi Romem, Chief Economist at BuildZoom

13. Igniting Curiosity by Benjamin Solomon, Director of Creative Development at Vox

14. Lessons from a Technology Journalist by Kim-Mai Cutler, Partner at Initialized Capital

15. The Paradox of Great Content by Rand Fishkin, Founder of Moz

Audience Growth: The Little Things that Work by Sam Parr, Founder of The Hustle

• The Hustle is an 18-month-old media company. It’s like WSJ meets Daily Show. The company is profitable and has a team of 18.

• The main product is a newsletter that has 500k subscribers. The website is just an email signup page.

• How did the Hustle grow? By doing small things very well.

  1. Everything is content. Creative copy is optimized for everything, including pop-ups (which convert at >5% for The Hustle vs. <1% industry average), welcome emails, Thank You pages, and so on. Writing fun, memorable copy for all of these “little things” keeps people much more engaged. Your email subscriber flow is Sign Up Page -> Welcome Email -> Thank You Page, and you should make those three pieces of content exceptional.

  2. Do unscalable things like replying to reader emails. The Hustle gets about 100-1000 emails per day. Along with making users happy, replying to emails also moves you from Gmail’s Spam or Promotions tabs into the regular tab. “It’s almost like SEO, but for email.”

  3. Create a community. The Hustle has 4k ambassadors, who invite others to The Hustle. (Sam recommended checking out this article as a good example of implementing a referral program.) Every few weeks, subscribers get an email with a referral URL, and if they share it with people then they get gifts like stickers, t-shirts, and hoodies. There’s also an online private FB group for Hustle ambassadors where the company asks them for feedback on new ideas. Each Hustle ambassador refers an average of twenty-seven people.

How to Scale a Content Team by Maggie Leung, VP of Content at NerdWallet

• NerdWallet helps people with personal finance decisions (credit cards, taxes, student loans, etc). Maggie previously worked at WSJ and CNN.

• NerdWallet’s mission: provide clarity for all of life’s financial decisions. Their main product is content and they have $100m+ in annual revenue.

• The content team is 100 people, mostly remote. Being remote helps ensure that they have a diverse team – i.e. not everyone lives in the Silicon Valley bubble.

• Before building a content team, you need to be crystal clear on your business. Specifically, you need to understand how content might drive results for you.

• Tactics for building (content) teams:

  • Specialist vs. generalist – most teams start with generalists, then shift to specialists once they figure out what they need specifically.

  • If you do heavy lifting at the beginning, you’ll save lots of time later. For example, NerdWallet tests everyone – even very experienced people – and only hires the top 1%. That saves NerdWallet from a lot of trouble later on.

  • Avoid hiring people who you think can do today’s job but not what the job will look like next month. No one wants to hire people only to have to fire them shortly thereafter.

  • Look for people who can adapt, not people who refuse to step out of their comfort zone.

  • Up-level quality: every new hire has to be better than at least 50% of the existing team.

  • Things you can’t teach: intellectual curiosity, team player/not an asshole, someone who can deal with unexpected challenges instead of shying away from them.

BuzzFeed: How We Data Science by Gilad Lotan, VP of Data Science at BuzzFeed

• BuzzFeed does investigative news, social news, entertainment content and shows, and now has a new news show on Twitter.

• Stats: 100s of millions of readers, content in 6 languages that’s published on 30 platforms.

• BuzzFeed has a culture of experimentation. They track core KPIs, then run experiments and compare new KPIs to baselines to see what works well. For example, BuzzFeed started making Tasty videos. Those performed well, BuzzFeed doubled down, and now Tasty is one of the most followed pages on Facebook (90m followers).

• When you see an experiment working, put more resources behind it.

• BuzzFeed cares about how its content affects the audience. They use ML to analyze content and comments in order to understand why people are sharing and what they get out of a piece of content.

• Optimization (personalized feeds and notifications, headline/image optimization, etc) is better than manipulation (buying traffic, buying followers, etc).

Why Data-Driven Companies Hire Economists by Jed Kolko, Chief Economist at Indeed

• Indeed Hiring Lab’s mission is to help lead the global labor market conversation, build the Indeed brand, and create a public service.

• The strategy is research and engagement using Indeed’s proprietary data + public data. The channels are blogging, social media, media requests, and speaking. The audiences are media, policymakers, and employers and job seekers.

• Why should you do public-facing research in-house?

  1. You can rapidly respond to news. For example during the last US presidential election, Indeed saw a big spike in US users searching for jobs in Canada, and they were able to write a post about that quickly.

  2. Broader public voice/thought leadership.

  3. More credibility for the company (this might be the most important benefit).

• PR is not the only benefit of a research team. Other benefits include:

  • Internal data quality (because you need good data for research).

  • Core product experience (sometimes insights from research teams affect product design).

  • Client/customer engagement (in Indeed’s case, they can help large employers understand what’s going on in the labor market).

  • Corporate strategy (research can suggest which parts of the business Indeed should grow, who they should try to acquire, etc).

• Build a public research team if:

  • You have unique data that tells us about the world.

  • You’re in an industry that’s of broad public interest (jobs, housing, etc).

  • You have a public-minded mission.

  • You are committed to maintaining high-quality data.

  • You aren’t in a high stakes public battle. (If you are in such a battle, people will be skeptical that your published results are impartial.)

Attract, Capture + Convert Customers by Shafqat Islam, CEO of NewsCred

• NewsCred powers content marketing for brands and works with companies like Microsoft and Bank of America.

• Content powers the customer experience. NewsCred believes content is the single most effective way to deliver more engaged, better-qualified buyers to digital experiences.

• Stats: 84% of people expect brands to product content, but 60% of content is poor or irrelevant.

• How to think about content: content has to be valuable to you (brand purpose) and valuable to your audience.

• Content has many forms: blog posts, videos, infographics, podcasts, social media, email newsletters, events (offline content), etc.

• Content creates value for your brand by driving marketing-influenced revenue (more leads, conversions, upsell and cross-sell) and increasing brand equity (awareness, loyalty, value of brand).

• Stat: the average NewsCred prospect reads 20 pieces of content before becoming a customer.

• NewsCred’s 3-step model for content: attract the right audience, capture/keep your audience, and convert the audience to a desired action.

• Attract

  • You need a strategy. Marketers with a documented content marketing strategy are 5x more likely to succeed. The strategy should cover your objectives, the resources you need, how you plan to curate and distribute content, and how you’ll measure and optimize performance.

  • Make sure you have a mission statement and define your target audience.

  • Identify topics your users are interested in by looking at keyword search volume and other signs of demand.

  • For content to be effective, brands must meet audiences where they are. But content on other sites must always lead back to an owned content hub (i.e. your website).

• Capture

  • NewsCred has a pop-up lightbox that triggers halfway down a page. They’ve found that lightboxes outperform other CTA placements by up to 10x. (NewsCred uses OptinMonster for the lightbox.)

• Convert

  • To convert people, NewsCred sends 4 newsletters per week. These have no promotional message other than links to high-value content. This makes people come back.

  • For NewsCred, the conversion rate for people nurtured via emails is 3x higher vs. those who just come to their website to sign up.

  • Articles should never be dead ends. There should always be a CTA or links to more content.

The Anatomy of a Viral Video by Grant Marek, Editorial Director at Chubbies Shorts

• Chubbies makes “clothing and content geared for the weekend.”

• They’ve produced 48 videos with 1m+ views, 13 with 10m+ views, 6 with 20m+ views.

• Viral videos don’t need celebrities or expensive production values. A lot of Chubbies’ videos take one day and <$1k to produce.

• Viral video strategies that have worked for Chubbies:

  1. Relatability. Your audience should be able to relate to the video content.

  2. An understanding of your audience: what they like, what they think is cool, and what they aspire to. If your content doesn’t appeal to your target audience, it’s useless. For example, Chubbies’ audience skews younger, so when they tried making videos that made cultural references from the early 90s and before, those videos didn’t perform well.

  3. Timeliness. People see the same content all of the time, and if you can make something timely to an event or a holiday then your content will stand out.

  4. Something unexpected. People are bombarded with content. If you want to produce something viral, it has to be unique and surprising.

  5. Make the videos fun. People burn out on serious content, news, and politics. Videos that are fun are an antidote, and much more likely to be shared.

• Finally, you need to know your metrics. One good metric for videos is the ratio of shares to views. A great goal would be 1 share per 100 views.

The POPSUGAR Formula by Lisa Sugar, Founder of POPSUGAR

• POPSUGAR’s mission: power the optimism and dreams of women around the world. They have 3.1b monthly content views, a global audience of 100m, and they’re the #1 lifestyle brand for women. The company started in 2005.

• The POPSUGAR formula for content is data (trends and sentiment) + aesthetic (beautiful and inspirational) + voice (positive and entertaining) + storytelling (smart, well-researched, fun).

• There’s so much potential content, so how do you filter it for your audience? First, understand the audience. POPSUGAR’s audience is women in their 20s. They’re optimistic and encountering many ‘firsts’, like living alone or getting married. The goal is to inspire and motivate these women, and the attitude of the content is positive, purposeful, playful, and passionate.

• POPSUGAR’s content engine is consists of many teams:

  • The Trending team (which identifies topics to write about)

  • The Social team (which creates social-only content)

  • The Audience team (which analyzes audience traffic and behavior)

  • Teams that generate content for specific topic areas (fashion, beauty, lifestyle, etc).

• The company publishes 1,000 articles and 10 original videos per week and sees about 1 million shares/day.

• POPSUGAR’s content workflow:

  1. Start with an idea (inspired by trends, events, brand relationship, etc)

  2. Apply a format to it (event coverage, social content, original feature, quiz, etc)

  3. Produce the content (headline, body, images (front page, social, and Pinterest-optimized), image and product sourcing, tags, SEO title, social text, etc)

  4. Content is reviewed by an editor and copy editor.

  5. Finally, content is distributed (scheduling, social marketing, push notifications, feeds, referral links and editorial partnerships, newsletter, etc)

Fireside Chat: Marketing for Startups with Michael Seibel, CEO of Y Combinator

• YC’s biggest marketing secret is to put you in the same room as lots of other people who are trying to grow because that makes you try to grow even harder.

• It’s better to have a few customers that love you instead of a lot that merely like you.

• Good cold email is a core strategy for 50%-60% of YC founders. “If you have a big problem and some young company emails you and offers to solve your problem, that could be the highlight of your day!”

• People think about launches as significant moments, but no one remembers the day that Amazon or Google launched. So your launch day isn’t important. Just get the product out ASAP and don’t wait for perfection.

• For launching, you should post on HackerNews, Product Hunt, TechCrunch. But none of those will make your company. Just try to get users.

• Good trick for early traction: get other friendly companies to use your product (e.g. companies in your coworking space, your accelerator’s batch, your investor’s portfolio, etc.)

• YC has a history of content, starting with Paul Graham’s essays. After Paul Graham retired, YC started writing blog posts and doing podcasts.

• “Content marketing keeps you relevant.”

• Tip for posting on HackerNews: don’t use marketing language. HN readers hate marketing language.

From Zero to Five Million Pageviews by Courtland Allen, Founder of Indie Hackers

• Courtland was always a CTO/coder and had never done content before. He shared tips on how he grew Indie Hackers into a high-traffic site.

• Courtland wanted to do something impactful, profitable, and fun. He decided to see what other people were doing. He looked at a lot of ideas, and that helped him figure out what he wanted to work on. He realized he liked finding stories of entrepreneurs, sprucing them up, and publishing them online.

• Good stories are transparent and honest.

• Courtland started interviewing entrepreneurs, but he was worried that his early interviews weren’t very good. He launched anyway, and got 100k pageviews on Day 1 from HackerNews, Reddit, and Product Hunt.

• Lesson 1: Go to where your audience already hangs out. After launch day, Courtland tried newsletters and other growth channels but none really worked. Finally, he went back to his initial channels (HackerNews/Reddit/Product Hunt), and those worked very well.

• Lesson 2: Systematize variety. Courtland tried to send the same questions to all interviewees, but the answers were often boring. So he started adding tips to questions – up to 20 tips per question on how to write better answers. After a few months, Courtland could get a good interview without any follow-up questions.

• Lesson 3: Content can look like anything. It can be blogs, newsletters, a recipe list, a spreadsheet of tourist attractions, etc. A good goal is that your content should be good enough to be its own product or startup. In fact, a lot of startup advice applies to content, too: work on (write about) what you know; launch (publish) quickly; do things that don’t scale (learn as you go); talk to users (readers); make something people want; etc.

Connecting with Journalists by Ricky Yean, CEO of Upbeat

• News reporting is shifting and there’s now less investigative reporting.

• More and more, reporters are generalists who have to write about many different topics that they’re not super deep on. This is stressful for reporters.

• Reporters often get 100s or 1000s of emails per day. Email pitches are not manageable. Competition for attention is intense.

  • Tip #1: if you want a journalist’s attention, you have to be the journalist: offer multiple sources, fact check, package insights, etc. Your email subject line should look like a headline, you should create high-quality graphics, photography, data, video, etc. Media organizations have fewer and fewer resources to do these things themselves. You can also de-risk pitches by posting on Reddit/HackerNews and showing that they’re getting traffic.

  • Tip #2: fight commoditized information flow by becoming proprietary. Use data/content that no one else has. Examples of good content that leverages proprietary data: Priceonomics/Earnest, Cardiogram.

  • Tip #3: PR pitching is unpredictable, so don’t be discouraged. Take more shots on goal.

• BuzzSumo is a good tool to help de-risk the popularity of a specific topic.

Hard-to-Make Data-Driven Content by Amber Thomas and Matt Daniels of Polygraph

• Polygraph produces pudding.cool, which is a journal for visual essays.

• Creating an evergreen piece of content with lots of data and visuals is very hard, and Polygraph has developed an editorial process to help mitigate common risks (e.g. overly ambitious projects) and encourage good practices like collaboration and sticking to deadlines.

• The process is 5 steps:

  1. Idea generation – the team generates ideas all of the time. Once a week, a member of the team will pick 3 ideas and present them to the team for Q&A and reactions. At this stage, ideas are just concept stage. All ideas discussed at these meetings are added to an idea backlog so that anyone can review past ideas at any time.

  2. Idea selection – there are two requirements for deciding whether to pursue an idea: 1) would a visual data-heavy essay answer a publicly divisive question or improve public discourse in some way? and 2) is the idea something the author passionate about (litmus test: would they work on this story for 1-2 months even if they weren’t being paid?)

  3. The pitch – a pitch helps articulate and flesh out the idea. At Polygraph, the team picks a story angle, a headline (that helps frame what you want the readers to get from your story), and does some preliminary data collection to validate that data is available and can support the desired story.

  4. Pre-production – the team finds a project lead who is accountable to push the project through to the end. They also make a rough outline of the story and scope how much work the story will take in terms of data and graphics and time. At this point, the team is fully committed to making the story and adds it to the editorial calendar.

  5. Production – the team starts working on the full story until it’s done. Tips for this part of the process are to do weekly check-ins with the editor, to vet the project with people outside of your organization, and to have a weekly meeting where people show what they’re working on, so that they can get feedback from others on the team.

On a completely unrelated note, I’ve spent hours on Polygraph’s visualization of evolving music tastes. It’s amazing.

Content that Caters to Multiple Tiers of Media by Issi Romem, Chief Economist at BuildZoom

• BuildZoom helps people find contractors.

• Why does BuildZoom produce content? 1) brand awareness, 2) SEO improves as you become an authority in your space, 3) qualifying web traffic (helps people associate your name with your industry).

• What kind of content does BuildZoom produce? Guides for things like how much roofs cost, news-relevant posts (e.g. cost of building Apple’s campus), and other analyses (e.g. curbing urban sprawl in the US).

• Credibility is very important when talking to reporters because their attention is spread so thin. How do you build credibility?

  1. Do great research.

  2. Seek external validation. Work with independent/academic coauthors if possible.

  3. Avoid topics with vested interest. For example, if you’re Airbnb, then publishing studies about how you help the housing market will hurt your credibility.

  4. Be transparent. It’s okay to make assumptions, but be clear about them. If you aren’t comfortable sharing your assumptions, maybe you shouldn’t share your piece.

• Posts can have different levels of appeal: national, national + some metros, etc. Examples: a study about national trends is unlikely to appeal to individual metros; a post about a single metro probably only appeals to that metro; and a ranking of metros nationwide mostly appeals to top and bottom ranked metros. If you want to appeal to every metro you’ll probably need deep content for each locale.

Igniting Curiosity by Benjamin Solomon, Director of Creative Development at Vox

• Curiosity is why we click, view, and opt-in.

• Some content is shallow: you’re curious, do a quick scan of the content, and you’re satisfied. Other content is deeper, and the more you read the more you want to learn.

• Vox’s learnings for building deeper content:

  • Respect your audience. Don’t just focus on content length or metrics or how many people click on something. Data can help you optimize, but fundamentally you need to understand your audience. If you create content that your audience is interested in, they will want to consume it and engage more deeply.

  • Foster obsessive creators. Let creators tell stories they’re authentically passionate about. Vox just launched Explainer Studio: when you view their content, you can see the explainer really puts themselves into each video. It’s not just a job for them, it’s something they’re passionate about. “Don’t think of content as a series of assets and think of it as human.”

  • Champion overdelivery. It’s not about giving the audience what they wanted, it’s about giving them so much more that they leave your content wanting to immerse themselves in a topic.

• “Focus on the story, not the shape. If you’re focusing on the length, you’re focusing on the wrong thing.”

Lessons from a Technology Journalist by Kim-Mai Cutler, Partner at Initialized Capital

• Kim worked at TechCrunch (where people could post short articles on whatever they wanted), WSJ (deep, research-based articles), and other publications.

• The press is not a deus ex machina.

• You have to re-iterate your goals before every press cycle.

• What are your goals for PR? Do you want to launch a product and reach customers? Hire or attract talent? Attract investors? Those could be very different stories to emphasize.

• Build long-term relationships w/reporters:

  • Do quarterly coffees or off-the-record meet-and-greets with reporters and influencers.

  • Aim for regular press cycles of 6-12 weeks so you stay top of mind with customers/investors/etc.

  • Have a longer term narrative arc about your vision that spans over your regular press cycles.

• When you pitch something to a reporter, make it custom to them. Base it on what they’ve written in the last 3-6 months + the philosophy of their publication and how that publication views the world.

• Story typologies, approximately sorted from easy-to-pitch to hard-to-pitch:

  • Product launch.

  • Funding round (used to be considered more important a decade ago, now less exciting).

  • Acquisition story.

  • Notable key hire you poached from someone else.

  • Key business partnership.

  • Momentum (i.e. “XYZ accelerates expansion into ___.” Numbers help here with stories like this, and this is a good way to set up fundraising).

  • Trend piece. Here the relationship with a reporter matters because it’s hard to become part of a piece on some trend. You already need to be in the reporter’s mindshare.

  • Feature story.

  • CEO/business profile. Even if you get a publication to do a profile on you or your business, the timing is tricky as these stories are often released a few months to a year after your conversations with the reporter.

• The press landscape has changed. More than ever, the top 5 tech companies (Google et al.) suck up most of the attention. Despite there being lots of tech journalists, it’s harder for small companies to get coverage.

The Paradox of Great Content by Rand Fishkin, Founder of Moz

(The deck for this presentation is here.)

• Search traffic matters a lot for companies, and content is a great way to build search traffic.

• Rand did an analysis of ninety-nine pieces of remarkable content around the web. Rand calls these “10x content.”

• On average, each of these 10x pieces has 900 links, 3k ranking keywords, and 80k Facebook shares.

• 8 big mistakes highlighted by separating the 10x pieces that perform vs. those that don’t:

  1. Lack of keyword research and targeting. If you optimize the title and content for popular keywords, you’ll get way more search traffic.

  2. Failure to optimize for SERP click-through rates. If you structure content in a certain way, like using tables instead of images of tables, that works much better in SERPs and can even get you a featured snippet on Google. You can get that snippet at the top of the page even if your page is no the top-ranked search result.

  3. Poor Domain/URL segmentation. In terms of SEO, blog.website.com performs much worse than website.com/blog. Moving to a subfolder helps the blog, and also the whole domain which now benefits from the blog’s SEO juice.

  4. Building on someone else’s platform. Don’t put incredible content on Medium and not your own website. It’s fine to put good content on Medium and your site, but don’t only put it on Medium. One tip is to publish on your site and use Medium’s import tool. Medium will even give your post a rel="canonical" link, which is very nice of them. “Use 3rd party platforms for distribution, not the primary home.”

  5. Lack of basic accessibility. Your site content should have shareable internal links and be easy to crawl, otherwise Google won’t be able to find your content – and your ability to rank for keywords will suffer.

  6. Delightful but not useful. Users will link to things that are useful, but not as much to things that are fun but not useful. That means your content won’t rank highly for keywords. Make sure your content is useful.

  7. Ignoring searchers’ intent. If someone does a search to learn something and your page tries to get them to buy something, they will bounce quickly and Google will learn from that and lower your search ranking.

  8. Doesn’t create an incentive to create a link. Rand gave the example of a tool that lets visitors compare data on two cities. The tool output was interesting but didn’t create an incentive to link to the content, so it had very few external links despite having thousands of upvotes on Reddit.

Q&A with Rand:

• Q: Google is taking more and more of Page 1 surface area. How do you keep search traffic high?

  A: 1) Try to own the featured search snippet. 2) If a keyword search can be answered so briefly that people stop clicking on search results, then pursue more interesting/complex keywords instead.

• Q: How do you balance headline quality/catchiness vs. SEO?

  A: Usually, if there’s deep interest in a subject, adding relevant keywords to a headline will increase interest. If there’s a 2-3 word phrase that’s important to you, including it in your title usually helps with SEO and clickability. It’s more challenging if the keyword phrase is long.

Well, that’s it! If you liked these conference notes, you might also enjoy my notes from SalesConf in 2014.

“All dollars are equal, but some dollars are more equal than others.”

- George Orwell’s imaginary capitalist twin

In my college economics class, I was taught that money is fungible because there’s no difference between one dollar and another dollar. Over the last five years as a VC, I’ve learned that that’s not exactly true and that some dollars are much more valuable than others to a business.

From a startup’s perspective, there are three common reasons why the value of a dollar can vary:

1. Money has different marginal value to a customer depending on where it falls on their balance sheet.

2. Some types of revenue require a lot more effort and resources to earn than other types.

3. Timing matters, and a dollar today is worth more than a dollar next year.

This post offers several concrete examples of these principles. The goal is to encourage thinking about revenue, costs, and value from different perspectives.

1. Cost vs. Revenue

Products create value by helping customers increase their revenue or decrease their costs. However, revenue and costs have very different structures.

Example: one of your customers is Etsy, which charges its users 7% for each transaction. If you help Etsy sell $1000 in extra merchandise, you’re creating $70 of value for them. If you save Etsy $1000 in costs, you’re creating $1000 in value for them. You could easily charge Etsy $100 if you reduce their costs by $1000, but you couldn’t charge them $100 if you help them sell an additional $1000 of products.

Lesson: a dollar of cost savings is worth one dollar to the customer, but a dollar of extra revenue is usually worth dimes or pennies (depending on the customer’s profit margin).

Exercise: think about how to maximize your product’s value to a customer while taking their margins into account.

2. The Principal-Agent Problem

According to Wikipedia:

The principal-agent problem occurs when one person or entity (the “agent”) is able to make decisions on behalf of, or that impact, another person or entity: the “principal”. This dilemma exists in circumstances where agents are motivated to act in their own best interests, which are contrary to those of their principals, and is an example of moral hazard.

Whenever the person buying a product is not the person who uses it, the principal-agent problem arises. An agent values things based on his/her own incentives, not the principal’s incentives.

Example: a customer’s VP of Engineering has ten direct reports and feels very overworked. If a product saves each engineer 1 hour every month but creates 2 hours of work for the VP, that will be a hard sell if the VP is the buyer. But if the product saves each engineer 30 minutes while also saving the VP three hours, that’ll be a much easier sell. In each case, the customer’s company saves 8 hours of engineering time per month (1 x 10 - 2 = 0.5 x 10 + 3 = 8), but the second product will be much easier to sell than the first.

Lesson: a dollar of value for the person buying your product will be worth much more to them than a dollar of value for someone else in their organization.

Exercise: think about how to create value for your buyer – especially if they are not the main user of your product.

3. Existing Expense vs. New Expense

Companies and teams often have existing budgets for common expenses. It’s easier to sell something if it fits into an existing budget.

Example: a company has a $500/mo learning budget for each employee. If you’re selling training software, then the company can pay you out of its learning budget. But if you’re selling something new, like “employee happiness monitoring” (EHM), then there’s no budget for that. Even if the company wants your EHM product, it’ll be hard to find a concrete buyer because no one has a budget for EHM.

Lesson: it’s easier to take a dollar from an existing budget than to take a dollar from a budget that doesn’t exist yet.

Exercise: if your product is creating a new spending category, think about whether there’s a way to reframe the product or adapt it to an existing category so that it’s easier to fit into an existing budget.

4. Above vs. Below Discretionary Spending Limits

For lower-priced products, customers often have discretionary spending limits. Products priced under these limits are much easier to sell.

Example: a customer’s engineers are allowed to spend $500/mo on whatever they want without their manager’s approval. If your engineering tool costs $499, then you’ll be able to sell it quickly. But if you raise the price to $501, then your sales cycles and conversion rates will worsen. In this case, the 501st dollar is very different from the 500th dollar.

Lesson: a dollar from someone’s discretionary spending limit is much easier to capture than a dollar that requires layers of approvals.

Exercise: figure out your customers’ spending limits. What kind of additional approvals or effort are required as you exceed each of those limits? How does that affect your product and pricing tiers?

5. Selling Services vs. Customized Products vs Off-the-shelf Products

Different types of revenue have very different scaling characteristics. Revenue that requires human labor is worse than revenue that doesn’t because the latter is much easier to scale.

Example:

• Product #1: you’re selling consulting services. Each employee can generate $300k in annual revenue.

• Product #2: you’re selling software that requires some initial customization by an employee. Each employee can handle customization for $2m worth of annual revenue.

• Product #3: you’re selling software that doesn’t require any labor. You can scale up very quickly without adding employees.

The difference between these alternatives is clear: if your startup is presented with a $30m consulting opportunity tomorrow, there’s no way to take it because you’d have to hire 100 people overnight. If the $30m opportunity is for customized software, then you’d “only” have to hire 15 people quickly. That’s doable, but still hard. But if you close a deal for $30m in off-the-shelf software, you should be able to deliver that without a hitch. The less labor a dollar of revenue requires from your company, the more attractive that dollar will be.

Lesson: selling a dollar of software has much better margins and scalability attributes than selling a dollar of services.

Exercise: are there specific use cases or customer segments where your product requires less customization and manual labor? What can you do to grow your sales for those use cases/segments?

6. Selling to Many Stakeholders vs. One Stakeholder

Products with a single stakeholder are straightforward to sell. Products with many stakeholders are hard to sell because you have to make everyone happy – but different stakeholders will have different, often conflicting, incentives and preferences. This example is loosely related to the earlier principal-agent example.

Example: healthcare systems have a lot of parties whose incentives may not be aligned. There’s the patient, the doctor, the hospital, the insurer, the government, and so on. As a result, healthcare software can be notoriously difficult to sell because it’s hard to please everyone. Maybe a patient wants a service, but the service is time-consuming for the doctor or not something that insurers will pay for. Or maybe a doctor loves a productivity tool, but the hospital doesn’t want to pay for it. In each of these scenarios, the likelihood of a sale is very low.

Lesson: the more parties a product involves, the harder it will be to sell. On the other hand, the fewer people have to approve a dollar of spending, the more likely you are to see that dollar. (This is related to discretionary spending limits, where going just above a limit triggers additional stakeholders.)

Exercise: can you sell a variant of your product that involves fewer stakeholders?

7. Monthly vs. Upfront Payments

A dollar today is worth more than a dollar next year. If you are paid today, you can use the proceeds to cover costs or to grow your business. If you’re paid later, you have to figure out how to stay afloat until the revenue finally comes in.

Example: you sell your product for $60k/year and you have $50k/mo in expenses. If you can close a new customer every month and your customers prepay annually, then you’ll be profitable every single month. If customers pay $5k monthly instead of $60k upfront, then you’ll be losing money until you land your tenth customer in month #10.

Lesson: it’s best to collect revenue upfront so that your business has a better cash cushion.

Exercise: are there incentives you could provide, both to your sales/marketing team and to your customers, that would help you receive more revenue upfront?

8. Selling vs. Upselling

It’s typically easier to sell more products to an existing customer than to find a new customer. According to the Pacific Crest SaaS Survey, the median SaaS company spends $1.13 to acquire each dollar of annual recurring revenue, but only $0.27 to generate an extra dollar of recurring revenue from an existing customer.

Example: it might take Dropbox $50 of marketing spend to acquire a new user who pays them $100/year, but only $20 of marketing spend to get an existing user to upgrade from a $100/year plan to a $200/year premium plan. Dropbox earns an extra $100 in both cases but pays a lot less to acquire that revenue from an existing customer.

Lesson: making a dollar from an existing customer is usually easier and cheaper than finding a new customer who will pay you a dollar.

Exercise: if you’re mostly focused on finding new customers, think about whether there are products or services that you could upsell to existing customers.

Conclusion

As you build out your business, consider the value of each dollar in the system to everyone involved. The better you can align your product with your buyers’ needs and constraints, the healthier your business will be. Your goal should be to find areas where your customers get more value and to find areas where you can expend less effort to produce that value.

I'm writing a post on what happens behind the scenes at a VC firm: partner meetings, funnel, deal flow sources, etc. Ideas for add'l topics?

— Leo Polovets (@lpolovets) March 23, 2017

Table of Contents

Part 1: Deals, Deals, Deals

A post on how VC funds source, analyze, and choose companies to invest in.

Part 2: Helping Founders and Time Allocation

A post on how investors interact with companies after investing, and how they allocate their time on a day-by-day basis.

Part 3: Fund Structure, Fundraising, Investor Relations, and FAQs [this post]

• Fund Structure

• Fund Economics

• Raising a Fund

• Investor Relations

• Venture Capital FAQs

  • Are there any ethical issues when deciding what companies to invest in?

  • What’s an investment thesis?

  • How frequently do VCs stick to their investing theses vs. investing opportunistically?

  • What are the advantages and disadvantages of an investing thesis?

  • What motivations does the VC have to invest or not invest?

Fund Structure

Venture funds have two principal parties: general partners (GPs) and limited partners (LPs).

LPs are the fund’s financial backers. These are the people whose capital is being invested. LPs can range from university endowments to pension funds to wealthy individuals.

GPs are the fund’s day-to-day managers. These are the people who make startup investments. They can be thought of as the middlemen that connect LPs’ capital to the founders who need funding for their startups. On top of allocating capital, good GPs usually try to provide value to founders in the form of advice or introductions or services.

Funds typically last for ten years. The first portion of that decade – typically 2-4 years – is an active investing period where funds make new investments. Most funds make 20-40 investments during this period, with a typical investment being 1-2% of the fund’s total capital. The remainder of the ten years is more passive. During this period, funds can’t make new investments, but they can follow-on and invest their remaining capital into existing portfolio companies’ subsequent funding rounds.

Most seed funds reserve a portion of their fund for follow-on investments. This is the portion of a fund allocated to follow-on investments rather than new investments. For example, a $60m fund with a 2:1 reserve ratio will use $20m for initial startup investments, and save $40m for future investments in those companies.

The mission of a venture fund is to invest in startups, have those investments appreciate over time, and have those startups exit within the fund’s ten-year lifetime. (An exit would be an acquisition or an IPO.) If some companies are still operating after ten years and need more time before an exit, then the ten-year period can be extended by 1-2 years with LP approval. When the fund is wrapping up, the GPs try to liquidate all of their remaining investment stakes.

Fund Economics

Like many hedge funds, a typical VC fund has a “2 and 20” fee structure. This means 2% of the fund is charged as a management fee each year, and the fund’s GPs and employees split 20% of the profits they generate. The profit-sharing portion is usually referred to as “carried interest” or “carry.”

Fun fact: according to Wikipedia, “The origin of carried interest can be traced to the 16th century, when European ships were crossing to Asia and the Americas. The captain of the ship would take a 20% share of the profit from the carried goods, to pay for the transport and the risk of sailing over oceans.”

The 2% annual management fee usually tapers off after a few years. For example, the fee might be 2% per year for 4 years (the active investing period), and then it goes down by 0.25% per year for the remaining 6 years. This fee is meant to cover GP salaries, employee salaries, legal and accounting fees, office space, and additional costs like event planning or consultant services. As you might expect, funds often increase in size over time because that means more management fees, more employees, higher salaries, and bigger budgets to help stand out to founders. (Ah, incentives.)

The way carry works is that all exit proceeds are returned to LPs until their capital is paid back, and then they get 80 cents on the dollar for future exits. For example, if a $50m fund has $20m exits in years 6, 7, and 8, then the first two $20m exits will be redistributed directly to LPs. The first half of the final $20m exit will also be redistributed directly to LPs, and the second half will be redistributed 80/20 between LPs and GPs. Future exits in the fund will be redistributed 80/20.

On a side note, if you thought employee option vesting was long, carry vesting is usually much longer. A typical vesting schedule might last for 8 years, and also include a clawback of 1 or 2 years of carry if a general partner leaves to join another competitor fund. This is why GPs rarely move from one fund to another.

Funds are usually “stacked,” which means that once a fund is mostly done with its active investing period, the GPs begin raising their next fund. That means 7 years after starting, the GPs might be finishing year 7 of Fund 1, year 4 of Fund 2, and year 1 of Fund 3.

Raising a Fund

Fundraising usually happens every 2-4 years for most funds, in conjunction with active investing periods. The goal is to have one fund’s active period begin immediately after the previous fund’s active period ends.

The fundraising process for venture funds has many parallels to the fundraising process for startups. For young funds, a fundraising cycle can take the better part of a year, or even longer. For established, successful funds, fundraising might take weeks or even days.

Fundraising involves building a pitch deck and having a compelling story about why your fund partnership is a good one to back. Once the pitch deck is ready, GPs will approach their existing LPs from previous funds (if applicable), wealthy individuals, family offices, fund-of-funds, and university endowments. These groups often allocate a small portion of their assets (e.g. 1%-10%) to alternative asset classes like venture capital and private equity.

When starting a new fund, GPs who don’t have strong track records in the industry usually focus on high net worth friends, family, colleagues, and friends-of-friends. As the GPs build a good track record over time, they become more appealing to fund-of-funds, endowments, and other “professional” LPs.

As with startup pitches, fund pitches can vary a lot. That said, the core things most LPs care about are the GPs’ hypothesis about promising areas to invest in, the fund’s structure (how many investments will it make? what’s the average check size? etc.), the GP team, and the team’s investing track record so far (if applicable).

If fundraising is taking a long time – and it usually is – GPs will often have a “first close” on 25%-50% of the target fund amount, and then a final close on the rest of the fund. The first close lets the GPs start making investments out of a small portion of the fund while they are still raising the remainder of the fund.

Investor Relations

At the core of investor relations is an annual LP meeting. This meeting takes anywhere from an hour or two to an entire day. The GPs talk about all of their active funds, specifically focusing on each fund’s financial performance, portfolio company highlights, and near-term plans. The GPs also discuss their outlook on the market, answer LP questions, and often invite a few portfolio founders to give short talks about their companies.

Aside from annual meetings, GPs typically write quarterly or biannual investor letters. These letters include summaries of how the overall portfolio is doing financially, as well as noteworthy portfolio company updates.

Finally, LPs and GPs are people, so aside from formal meetings and investor letters, they get together casually for occasional catch-up meetings. LPs who write bigger checks might ask to have a regularly scheduled meeting every few months.

Venture Capital FAQs

Before writing this post, I asked people on Twitter what questions they had about working in venture capital. Here are some of the questions that were brought up:

Q: Are there any ethical issues when deciding what companies to invest in?

A: Sometimes LP/GP agreements will spell out areas that can’t be invested in (e.g. no drug or gambling startups). Many (but not all) funds won’t invest in companies that are shady in some way (i.e. companies that exploit underprivileged segments, use dark patterns, etc). However, unless undesirable investment areas are spelled out in the LP/GP agreement, investments are made at the discretion of GPs. If a GP does make an investment that LPs strongly question, then there’s a high risk of those LPs not investing in the GP’s next fund.

Q: What’s an investment thesis?

A: An investment thesis describes the types of companies that a fund wants to invest in. It’s a public hypothesis about the types of companies that have lot of potential over the next 5-10 years. A few examples of theses:

• “Companies reimagining the future of work” (Bloomberg Beta)

• “Hardware companies” (Root Ventures)

• “Companies with sustainable competitive advantages” (Susa Ventures)

• “Companies applying deep tech to big industries” (Data Collective).

Q: How frequently do VCs stick to their investing theses vs. investing opportunistically?

A: Not every VC has an investing thesis – some invest opportunistically in any startup that they like. For the ones that have a thesis, most investments will fall into that thesis. The reason for this is perception. Let’s say a fund’s thesis is that SaaS companies are the most promising startup category, and half of the fund’s investments are SaaS companies while half are not. If all of the fund’s best investments over time turn out to be non-SaaS, future LPs will question the fund managers’ judgment.

That said, if a fund meets the next Facebook, they won’t want to pass on it just because it’s out-of-thesis. Putting all of this together, most investors will stick to their thesis “most” of the time (i.e. 75%-100%).

As a corollary, if you’re a founder and your startup falls outside of a fund’s public thesis, it’s probably wise to approach other funds first.

Q: What are the advantages and disadvantages of an investing thesis?

A: The advantages are that if you become known as being a good, knowledgeable investor in a certain area, then you’ll see a lot more investment opportunities in that area: founders will come to you because you’ve shown you understand and invest in their space; other investors will introduce you to relevant companies to get your expert opinion; industry conferences will invite you to be a speaker; and so on.

The main disadvantage of a thesis is that it becomes harder to invest in other areas. If you’re known as a great hardware investor, then it might not occur to Mark Zuckerberg to approach you because Facebook isn’t hardware. What’s worse, if Mark does approach you, it’s probably because all of the more well-suited funds already said no. Because you’ll be missing out on a lot of out-of-thesis startups, you better have a lot of conviction in your thesis being correct.

Q: What motivations does the VC have to invest or not invest?

A: VCs want to make investments because every time they finish investing out of one fund, they can start the next fund (with more management fees and more carry). This is counterbalanced by the desire to be patient and only make good investments. If a VC makes good investments, they’ll raise more funds and earn more carried interest; if they make poor investments, they’ll stop being able to raise funds and they won’t earn much carried interest.

Most funds spend 2-4 years actively investing. Funds longer than 4 years would require raising too much LP capital, and LPs would rather see results every few years before deciding whether to continue investing in a fund. Funds shorter than 2 years are taking on a lot of time diversification risk: if all of a fund’s investments were made in a year that turns out to be a terrible year for startups, then LPs will lose their money. Spreading investments across 2-4 years mitigates this risk.

To slice this question differently, the main reason most investments happen is that a VC genuinely believes that a company has a shot at a $1b+ exit someday. However, sometimes there are other considerations:

• Sometimes GPs want a recognizable logo in their portfolio (e.g. one might invest in a late round of Uber or Pinterest just to say they are invested in those companies).

• Sometimes investing in a company will provide expertise or connections to good investment opportunities later on. For example, a fund might make several small investments into an emerging space like AgTech or cryptocurrencies, with the goal of building up enough knowledge to make larger investments in those areas in the future.

• Sometimes investing is an option bet: the goal is less about returns on the current investment and more about the option to invest a lot more later on. For example, sometimes Series A funds will invest $100k into a seed stage company. Those funds don’t really care about the $100k. What they care about is having the inside track on a company so that if it starts doing well, they can invest $5m or $10m in a Series A round.

If you have other questions about venture capital, please let me know on Twitter. I’ll add popular topics to this FAQ or address them in future blog posts.

Table of Contents

Part 1: Deals, Deals, Deals

A post on how VC funds source, analyze, and choose companies to invest in.

Part 2: Helping Founders and Time Allocation [this post]

• Basic Terms

• Why do VCs Help Portfolio Companies?

• How Much do VCs Help Portfolio Companies?

• Ways that VCs Can Help

• VC Time Allocation

  • Meetings

  • Email

  • Research

  • Events

Part 3: Fund Structure, Fundraising, Investor Relations, and FAQs

A post on the basic mechanics of how VC funds are structured and raised, how VCs interact with their own investors, and venture capital FAQs.

Basic Terms

Portfolio Company – a company that the fund has already invested in.

Limited Partners (or LPs) – a fund’s financial backers. These are the people whose capital is being invested. LPs can range from university endowments to pension funds to wealthy individuals.

Why do VCs Help Portfolio Companies?

Most good VCs try to help their portfolio companies for several reasons:

• If VCs are helpful, founders will recommend them to other founders.

• Being helpful can materially increase the value of a company – and by extension, the value of the VC’s investment. For example, let’s say my fund owns 10% of a SaaS startup. That type of company tends to trade at 5x revenue multiples in the public stock market (e.g. $200m in revenue -> $1b valuation). If I introduce that company to 10 customers that create $1.5m of additional revenue, then the paper value of my portfolio goes up by $750k (= 10% ownership x $1.5m revenue x 5.0 revenue multiple).

• On a more human note, most VCs prefer to think of themselves as company advisors rather than money managers. This often results in VCs wanting to help just so that they feel useful. (The downside here is that if a VC can’t be useful but tries anyway, they just get in the way.)

How Much do VCs Help Portfolio Companies?

Motivations aside, the level and frequency of engagement vary greatly from investor to investor. Here are the levels that I’ve observed:

• Uninvolved. The investor writes a check and then you never hear from them again.

• Passive. The investor doesn’t reach out proactively, but invites you to ask them for help when you need it. Sometimes these “let me know if I can help” offers are lip service, and sometimes they’re genuine.

• Semi-active. Most founders send monthly update emails, and these emails often have an Asks section that lists areas where investors and advisors can help. Semi-active investors will respond to those asks when they have something to offer. Here the founder is not asking for help one-on-one, but they still have to ask.

• Active. Active investors will reach out proactively when they have something they can offer (like a potential engineering candidate or an intro to a customer). They’re often thinking about ways to help in the background, even if founders aren’t explicitly asking them to.

• Very active. The next level of engagement typically involves a regular meeting cadence – sometimes as often as weekly, but more commonly every 3-6 weeks. At these meetings, which are almost like informal board meetings, founders will talk about how their business has been going for the last few weeks. As the VC listens, they’ll ask probing questions – both to understand the business better and to help the founder clarify their thinking – and suggest ideas or improvements or potential intros. For example, a founder might talk about how their recent experiments with Facebook ads aren’t going well. The VC will ask about the experiments, and based on the answers might offer an intro to a Facebook ads expert, or recommend a marketing channel that has worked better for other similar companies.

• Board level. Board level engagement is a more formal version of the “very active” level. The VC and founders often work together for multiple hours every month, and sometimes every week. For active board members, working with a single company might represent 5% or even 10% of their total work hours. This is why most later stage VCs top out at 8-10 simultaneous board seats.

While the level of involvement varies across investors, generally the more capital a VC is investing, the more involved they will be. That said, there are large funds that are very passive and small angel investors that are very active.

Ways that VCs Can Help

As I wrote ~3 years ago, good investors can help in many ways. In practice, the most common areas where VCs help are:

• Intros to prospective customers, job candidates, advisors, service providers, other investors, etc.

• Comparables and benchmarks for running your business, including compensation benchmarks, business benchmarks (e.g. typical churn levels or salesperson commission structures), and office space comps.

• Feedback on website copy, sales and marketing materials, product UI/UX, etc.

• General business advice, including tips for hiring execs, fundraising, choosing KPIs, and so on.

• Interviewing and helping to close key job candidates. Sometimes talking to a VC gives an on-the-fence VP the push they need to join a startup.

• High level strategic guidance. Experienced investors have watched 10+ or even 100+ companies grow (and struggle), and they can see early signs of challenges of opportunities. Sometimes the most valuable aid an investor can offer is explaining what’s likely to be around the corner in 6-12 months at a time when the founder is spending most their time thinking about tomorrow or next week.

As a meta-observation, founders have knowledge that’s deep but narrow while investors have knowledge that’s shallow but wide. Investors understand common patterns of success and failure across many types of businesses, but chances are they know very little about a specific business or industry. Founders on the other hand know a ton about their specific business, but haven’t had a chance to see what does and doesn’t work across many companies. As a result, VC advice is often most valuable when it’s applicable across many companies, but founders are probably wasting they’re time if they’re asking a VC for specific SEO tips or how to sell to plant managers in the Oil and Gas industry.

VC Time Allocation

As a VC, I spend most of my time on meetings, email, and research. I analyzed data from Google Calendar and Gmail on where my time went over the last few months, and the results are described in the remainder of this post.

Meetings

VCs meet with founders, service providers, LPs, other VCs, job candidates for portfolio companies, and many other parties. To give a sense of how those meetings break down, I categorized and logged all of my meetings for May, June, and July of this year.

Here’s how most of the days feel:

Here’s the average number of hours of meetings I had by weekday:

My office is in San Francisco, but I live 25 miles south in Redwood City. If you’re wondering about why Tuesdays and Fridays seem so much busier than others days, it’s because those are the days I usually spend in SF. Somewhere in here is an observation about how most Silicon Valley startups are moving out of the peninsula and into SF.

Here’s the distribution of hours of meetings per day:

Most days had 3-6 hours of meetings, although there were outliers on both sides.

Here are the types of meetings I had:

Clockwise from the top right, the categories are:

• 1:1 diligence – one-on-one pitch meetings (or calls) with founders of potential investments (3.5 hours/week)

• Group diligence – full team pitch meetings (or calls) with founders of potential investments (2.75 hours/week)

• Misc diligence – other diligence-related meetings for prospective investments: founder reference calls, customer reference calls, legal work, and so on. (1.25 hours/week)

• VC networking – meetings with other VCs. These meetings are with earlier stage investors (whose companies I might invest in in the future), similar stage investors (whom I can partner with on new investments), and later stage investors (whom I can introduce to my existing portfolio companies). (2.5 hours/week)

• Partner meetings – internal meetings where the fund’s partners discuss ongoing action items: decisions and next steps on new investments, ideas for helping existing portfolio companies, hiring decisions, upcoming events, and so on. (1.75 hours/week)

• Feedback meetings – meetings with founders who are not currently fundraising, but are looking for feedback on their idea/business plan/pitch in preparation for a future fundraise. (2.75 hours/week)

• Fund operations – miscellaneous meetings related to running the fund: interviewing candidates, talking to service providers, event planning, team off-sites, etc. (1.5 hours/week)

• Portfolio help – meetings that help portfolio companies in some way: brainstorming sessions with founders, interviewing key job candidates on their behalf, helping them prepare for the next funding round, and so on. (4.25 hours/week)

• LP-related – meetings with current and potential investors for our fund. (1 hour/week)

• Special events – various dinners and demo days. (1.5 hours/week)

The chart shows that about 33% of my meeting time is spent on diligence, 22% is spent on networking with VCs and founders who are not fundraising, 19% is spent helping portfolio companies, 19% is spent on internal fund work (partner meetings, hiring, LP updates), and 7% is left over for special events. On average, I have about 23 hours of meetings every week.

Email

According to Gmail Meter, I receive 2,000 emails and send 1,000 emails in a typical month. If you assume it takes 1 minute to read a typical email and 2 minutes to write one, this represents about 15 hours of email time every week.

Research

Many of the meeting categories above require doing some work before and/or after each meeting. For example, I spend about 15 minutes preparing before every 1:1 diligence meeting. If the meeting goes well, I’ll spend 30-60 minutes doing research and typing up notes for the other partners in my fund. If the meeting doesn’t go well, I’ll spend 10-20 minutes writing a nice email to the founder that explains why I’m passing on the investment opportunity.

Events

The last part of my job revolves around occasional special events. These include things like Y Combinator’s biannual Demo Days, which are 3-day events where 100+ founders pitch their companies to investors. My schedule fills up for 1-2 weeks before and after each of these demo days.

My fund also puts together an annual LP day for our own investors, and an annual founder/investor retreat. Both of these take weeks of planning and preparation. These types of events only happen a few times a year, but when one is coming up it doubles my workload.

Conclusion

This post talked about how investors engage with portfolio companies and how VCs† spend their time. The next and final post in this series will talk about how venture funds are structured and how VCs work with their own investors.

† By “VCs” I mean “me,” because that’s the only VC that I have lots of data for.

Table of Contents

VCs have 3 principal jobs: picking startups to invest in, helping startups after investing, and raising capital for the fund. Each of these jobs will be covered in its own post.

Part 1: Deals, Deals, Deals [this post]

• Basic Terms

• Finding Startup Investments

• The Due Diligence Process

• Pitch Meeting Structure

• Partner Meeting Structure

• Making Investment Decisions

Part 2: Helping Founders and Time Allocation

A post on how investors interact with companies after investing, and how they allocate their time on a day-by-day basis.

Part 3: Fund Structure, Fundraising, Investor Relations, and FAQs

A post on the basic mechanics of how VC funds are structured and raised, how VCs interact with their own investors, and venture capital FAQs.

Basic Terms

Portfolio Company – a company that the fund has already invested in.

Deal – a startup investment opportunity. Common usages: “we did a drone deal last month” and “did you see that autonomous car deal from Boston?”

Finding Startup Investments

When looking for startups to invest in, most funds rely on many different sources:

• Accelerators and Incubators (Y Combinator, Techstars, Alchemist, Acceleprise, etc). Most of these accelerators have demo days every 3-6 months, and a typical demo day features 10-30 companies. Y Combinator is an exception, typically showcasing 100+ companies over two days.

• Earlier stage investors, like angels or pre-seed funds. These investors might get involved with a company 3-6 months before its seed round. Once the company makes a little bit of progress, its early investors will introduce it to seed funds.

• Other seed funds. A typical seed round in Silicon Valley might include 2-5 seed funds. Because no seed fund takes up the whole round, whenever one fund invests, they will usually introduce the founder to other good seed funds.

• Series A funds. Sometimes later stage funds see companies that are too early for them, so they forward those companies to seed funds that they like working with.

• Introductions from founders. These introductions might come from founders in an investor’s portfolio, or from founders whose companies the investor has previously passed on. These intros are an incentive for investors to be as nice and as helpful to founders as possible.

• Miscellaneous intros from service providers, friends, ex-colleagues, your parents’ neighbor’s best friend, etc. These intros are good, but often weaker than founder intros. “You should invest in Bob’s company because Bob is my son’s best friend” just isn’t that compelling.

• Cold inbound emails from founders. Investors often receive dozens or even hundreds of inbound pitches every month.

• Cold outbound emails to founders. Many VCs will actively reach out to founders. This commonly happens if an investor lacks a warm intro to a company that they’ve heard about, or if an investor is doing a deep dive into a sector and trying to talk to as many promising companies in that sector as they can.

The Due Diligence Process

In order to figure out which companies to invest in, most seed funds have a standard diligence process. A typical process will contain the following steps:

1. Screening. One of the partners sees a demo day pitch or gets a cold email or a warm intro offer to a company. The partner spends 0-15 minutes on research to decide if the company might be a good fit for the fund.

2. 1:1 pitch meeting. If the company might be a fit, the partner will do a 30-60 minute pitch meeting (either in person or via phone).

3. Research. If the 1:1 meeting goes well, the partner will do some more research and ask the founder additional questions – either over email or another 1:1 meeting/call.

4. Group meeting. If the research phase goes well, the founder is invited to do a pitch meeting with all of the fund’s partners.

5. Customer and founder references. For companies that do very well during a group meeting, the partners will do customer and founder reference calls. The purpose of customer calls is to get more insight into how customers view a product, how much they value it, how they’ve liked working with the startup in question, how they discovered the product in the first place, and so on.

The investment funnel for a single calendar year might look like this:

*2000* screened companies -> *500* 1:1 partner meetings -> *300* further research -> *150* group meetings -> *25* further diligence (customer/founder calls) -> *12* investments

Pitch Meeting Structure

Pitch meetings are usually 30-60 minutes. They tend to be shorter during the early phases of diligence and longer in later stages where more partners are involved.

A typical meeting might start with a few minutes of friendly chit-chat. After that, everyone introduces themselves, and then the founder walks through their pitch. Sometimes the walkthrough is a few minutes because everyone around the table has been doing homework and research for a while, and sometimes it’s 15-30 minutes when the material is new to everyone.

After the pitch, investors will ask a lot of questions about the startup. Their goal is to figure out if they might be looking at one of the top ~1% of companies they’ll see all year. Whether something might be in the top 1% is based on many things, and common topics of discussion include:

• The founding team: what are the founders’ backgrounds? How long they’ve known each other? What are their individual strengths and weaknesses?

• The problem being tackled: how severe is the problem (the old vitamin vs painkiller dichotomy)? How many people have it? How much would customers pay for the solution?

• The solution being proposed: how is it different from existing solutions? How much better is it? How long would it take to build and how much has been built already? Is there a demo? What are future directions for the product?

• Traction so far: are there users or customers? How many? How fast is usage growing? How engaged are the customers? Do they ever churn, and if so, why?

• Competitive landscape: who is the company competing with? How does it differentiate from its closest competitors? Is there a sustainable competitive advantage (i.e. moat) that ideally gets stronger over time?

• Fundraising plans: how much does the company want to raise? How will capital be used? How much time will the seed funding buy, and what kind of milestones could be reached during that period? Is the valuation for the round set, or is the company looking for a lead investor to set the valuation?

Finally, pitch meetings usually end with the founders having a chance to ask the VC questions. (The two most common questions that founders ask, by far: “How can you help beyond capital?” and “What’s your decision-making process and what are the next steps?”)

Partner Meeting Structure

Partner meetings are weekly multi-hour meetings where the bulk of a fund’s decision-making occurs. A typical partner meeting agenda could include:

• Updates on portfolio companies. Each partner discusses updates on companies they’ve met with or emailed with since the last partner meeting.

• Back office discussions. Partner meetings are a good venue for discussing upcoming hiring decisions, accounting or legal TODOs, fundraising plans, and so on.

• Industry analysis. Partners often talk about trends they’re noticing, observations about the funding climate, and other macro topics that are worth discussing with the entire group.

• Decisions on startups that are being diligenced. Partners will go one-by-one through the active companies in their deal-tracking tool (e.g. Affinity). For each company, there’s a decision made about whether to invest, continue doing diligence, or pass. This is the longest, most important part of partner meetings, and it will be discussed next.

Making Investment Decisions

A typical seed VC that’s exposed to 1000-2000 companies per year might only make 10-20 investments. As you might expect from these numbers, picking companies to invest in is probably the hardest part of being a VC.

Most Silicon Valley investors aspire to find companies that have the potential to reach $1b+ exits because those mega-exits are the source of the majority of venture returns. The challenge is that while a 50th percentile company looks very different from a 90th percentile company, and the 90th percentile is distinguishable from the 97th percentile, the top few percent of companies are all very good. It’s hard to to guess which of those top companies have a shot to exit for $1b or $5b (which would be amazing for a seed fund) and which ones are more likely to top out at $40m or $100m (which is still very good, but far from amazing for the fund).

To help make these educated guesses decisions, investors usually think about three attributes: the quality of the founding team, the size of the market, and the product itself. Each investor tends to weigh these attributes differently – i.e. would they prefer to invest in a stellar team with a good product or a good team with a stellar product? A stellar team and a stellar product would be ideal, but that’s (unsurprisingly) rare.

If a fund has multiple partners – and most do – then all of the partners’ votes for each investing decision need to be weighted. Here are different vote-weighting strategies that I’ve heard of:

• Every partner has to love a deal.

• Most partners have to love a deal.

• If one partner loves a deal, the investment happens regardless of what anyone else thinks.

Additionally, some funds allow partners to veto investments, others do not. Finally, funds where investments require buy-in from multiple partners sometimes grant each partner a “silver bullet,” which is the right to make a unilateral investment once every few years.

After all diligence for a company has been completed, the partners will discuss their thoughts at the partner meeting. Each person talks about what they like and dislike about an investment opportunity, and then offers their vote. After the votes are weighted, the final decision is made on whether to invest or to pass.

If the partners decide to pass, one of them will notify the founder over a phone call or via email. Ideally, the decision to pass is conveyed along with the reasons for passing. In reality, some investors give great feedback, others say something generic like “sorry, this is too early for us,” and others just go cold (which is a jerk move).

If the partners decide to invest, one of them will reach out to the founder and make an offer. Sometimes the founder might negotiate the terms, or ask to speak to a few of the VC’s portfolio founders as a reference check. After that’s done, the deal enters the legal paperwork phase, which can take anywhere from a day or two to a month or two.

Conclusion

This post talked about how funds source, analyze, and choose companies to invest in. The next post in this series will cover how investors try to help their portfolio companies and how they spend their time in general. To give a sense of the latter, the post will include a numeric breakdown of where I spent my time over a 2-month period where I tracked and categorized every meeting.

Slides

De-Risking Your Startup -- SaaStr 2017 Talk

Audio

Unfortunately, the talk was not videotaped. However, here is an audio recording:

If you want to follow along in the deck, here are the approximate slide transitions:

Slide  Start   End
  1    00:00  00:09
  2    00:09  00:46
  3    00:46  01:06
  4    01:06  01:10
  5    01:10  01:44
  6    01:44  01:59
  7    01:59  02:21
  8    02:21  03:03
  9    03:03  03:15
 10    03:15  03:43 
 11    03:43  04:17 
 12    04:17  04:40 
 13    04:40  05:05 
 14    05:05  05:13 
 15    05:13  05:33 
 16    05:33  05:54 
 17    05:54  05:58 
 18    05:58  06:09 
 19    06:09  07:17 
 20    07:17  07:34 
 21    07:34  07:41 
 22    07:41  07:56 
 23    07:56  09:12 
 24    09:12  09:35 
 25    09:35  09:40 
 26    09:40  09:49 
 27    09:49  10:03 
 28    10:03  10:33 
 29    10:33  11:38 
 30    11:38  12:23 
 31    12:23  12:50 
 32    12:50  15:36

Further reading

This talk was based on two posts that I’ve written about de-risking:

• Startups are Risk Bundles – An explanation of the relationship between risks and startup valuations.

• How to De-Risk a Startup – A list of ways to remove risk from your startup in various categories like product/market fit or recruiting ability.

“All models are wrong, but some are useful.”

- George E. P. Box

It’s very hard, maybe impossible, to create a mathematical model that describes how to build a successful company. Still, there’s value in trying to come up with different models and thinking about their implications. This post describe three possible models for calculating a startup’s value and the lessons that each model offers.

Model #1: A Product of Risk Multipliers

In How to De-Risk A Startup, I described nine major risks that most startups face, as well as how investors might evaluate each of those risks:

1. Product/Market Fit Risk (is your product something that people want?)

2. Product Quality Risk (can you build a great product?)

3. Team Risk (do you have a great team for achieving your vision?)

4. Recruiting Risk (are you able to grow your team effectively with strong talent?)

5. Sales Risk (can you and your team sell your product effectively?)

6. Market Risk (is your target market large enough to support a huge company?)

7. Funding Risk (do you have enough capital to hit milestones that will either let you raise more capital on better terms, or get you to a place where you no longer need to raise capital?)

8. Short-Term Competition Risk (are you differentiated enough from existing competitors?)

9. Long-Term Competition Risk (will you be able to fend off strong, well-funded competitors in the long-run?)

For each of these categories, we can assign a ‘risk multiplier’ between 0.0 and 1.0. The more risk there is in a category, the lower the multiplier and the lower the expected value of a company. For example, a company where the team is inexperienced might have a Team Risk Multiplier of 0.2, while an experienced team would have a Team Risk Multiplier of 0.8.

Combining these multipliers, the expected value of a startup might be something like:

(The ∝ sign means ‘is proportional to.’ This formula is not an equality because it ignores factors like the role of luck, the revenue multiples that companies trade at for a given sector, etc.)

According to this model, the expected value of a company is largely determined by its biggest weaknesses.

Model #1: Implications

1) Startups are like triathlons

The model above suggests that a company’s top priority should be reducing its biggest risks. For example, if your startup’s risk multipliers are 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, and 0.05, then you should focus on the 0.05. Moving a risk multiplier from 0.75 to 0.90 improves your startup’s expected value by 20%, but moving a risk multiplier from 0.05 to 0.25 improves your startup’s expected value by 400%.

Under this model, startups are like triathlons. Triathlons consist of three events: a swim, a bike ride, and a run. To be a world-class triathlete, you have be good at all three sports. If you’re an amazing runner and cyclist but don’t know how to swim, then swimming is the only thing you should be practicing. It won’t matter if you can bike or run 10% faster if you can’t finish the first portion of the race.

2) You have to get most things right

The corollary to the previous implication is that you have to get most things right to build a huge company. If you’re going after a $10b market, and you have 10 risk multipliers each of which is currently 0.3, then the expected value of your company will be close to $0. Bringing each risk multiplier up to 0.5 gets you to an expected value of $10m. But your company won’t actually get into the $1b+ territory until your risk multipliers are in the 0.8 or 0.9 range and you’ve de-risked almost every aspect of your business.

3) Most investors look for easy reasons to say ‘no’

This model hints at why investors often say ‘no’ to companies quickly without doing much diligence. If the value of one or two of your risk multipliers is very close to 0, then many investors will pass because the 0’s drown out everything else. For example, if your funding risk multiplier is 0.01 (maybe you need to raise $50m upfront to implement your business plan, which is nearly impossible), then it hardly matters if other risk multipliers are 0.2s or 0.5s or 0.9s.

4) Market size is a critical factor

It doesn’t matter how high or low your risk multipliers are, if you’re going after a $10m market then you’re not going to end up with a billion-dollar company.

5) Early stage startup valuations are based on potential

This model explains how two founders can raise money at a $5m valuation on the day their company is founded. While a late stage company’s valuation is based on revenue and growth rate, an early stage company’s valuation is based on its potential best-case long-term valuation multiplied by the best-case scenario’s likelihood. So while the two founders haven’t built anything before incorporating, the combination of their idea, skills, and target market might suggest they have a 1% shot of building a $3b company over 10 years. That means the company has an expected value of $30m, and that justifies a $5m valuation today.

Model #2: A product of categorized risk multipliers

A small tweak to the first model is to group the risks into categories and to treat each category a little differently. The nine risks mentioned at the beginning of this post can be divided into three groups:

1. Your company’s current state (product/market fit, product quality, etc.)

2. Your company’s ability to execute and to rapidly improve its current state (the strength of your team, your ability to recruit, whether the company culture suppresses talent or amplifies it, etc.)

3. The competitiveness in your target market (short-term + long-term competition risk)

Here’s a (somewhat arbitrary) alternative model that reflects how these groupings might determine a startup’s value:

Why is execution ability given more weight than a company’s current state? Because a company’s current state is significantly easier to change than its execution ability. A strong team can quickly iterate on an okay product and make it great, but a weak team will have a hard time even if they start with an initially better product – and they’ll also have a hard time making the team stronger through recruiting because strong players don’t like joining teams full of weak players.

In mathematical terms, it’s better to have a CurrentState of 0.3 and a team that can improve that by 0.04 every month, than to have a CurrentState of 0.5 and a team that can improve that by 0.01 every month. The model reflects this.

Model #2: Implications

1) You should pursue a monopoly, a large market, or both.

The best way to maximize the rightmost term in this model is to find a market where MarketSize is large or CompetitionFactor is small. Ideally both. If you’re pursuing a small or mid sized market with a lot of competition then your chances of building a huge company are slim to none. On the other hand, heavy competition makes it very hard to stand out to customers, even if your company has a superior product and team.

2) People are hugely important in the early days of a company

At inception, a company’s CurrentState is close to 0 while its ExecutionAbility might be much higher than 0. The better the team and the faster it can execute, the faster the company’s current state will improve. As a result, the quality of a founding team is a very large factor in determining a startup’s seed stage valuation. This is why seasoned execs and successful serial founders often command higher seed valuations than first-time founders.

3) There’s more than one way to increase expected value

There are four ways to make your company more valuable:

1. Level up your current state: improve product/market fit, improve the product quality, build up your revenues, etc.

2. Level up your ability to execute: hire great people, design a culture that makes people more productive, offer training, build up automation, etc.

3. Look for adjacent markets that increase market size: other customers who would buy the same software as your existing customers, other software that your existing customers would buy, etc.

4. Improve your differentiation: the better you can differentiate yourself the lower your CompetitorFactor drops. For example, the car industry is very competitive and there are many car manufacturers vying for market share. However, Telsa has managed to differentiate itself so much that it has very little direct competition while still having a chance to capture a large chunk of the entire auto market. Companies that reach this level of differentiation have a lot more potential value and can effectively operate like monopolies.

Model #3: Achieving success one milestone at a time

In this model, the key focus is on reaching a progression of major company milestones. These milestones might be things like building an MVP, achieving product/market fit, hitting your first $1m in revenue, or finding a scalable customer acquisition channel. Each milestone increases the value of the company, and each depends on previous milestones – that is, you can’t find product/market fit without building a product, and you can’t find a scalable marketing channel until you have product/market fit.

Model #3: Implications

1) Most of your focus should be on the next milestone

If you just reached Milestone #3, you should start focusing on Milestone #4, not Milestone #7. For example, if you just built an MVP, don’t worry about hiring the CFO that you’ll need if you go public some day, focus on finding a good user acquisition channel instead.

2) Lay the groundwork for future milestones

Under this model, your company’s value won’t increase when you skip milestones. However, if you know your next few goals and know what’s needed to achieve them, you can set yourself up for success by laying some groundwork ahead of time. For example, if your current goal is to hit $500k in annual revenue and the next goal is to hire a VP of Sales, then on the one hand most of your focus should be on revenue growth, but on the other hand finding a good VP takes time, so starting to do interviews now is a good idea.

The value of models

This post presented a few (somewhat arbitrary) mental models for valuing startups. None of these models is ideal**, but each of them has interesting implications for how to maximize the long-term value of a company. While no early stage investor or founder should be explicitly calculating a startup’s value using these formulas, the implications are still relevant: you should mostly focus on near term milestones and not things you won’t need for several years, you should beef up your company’s biggest weaknesses, you should prioritize building a strong team because that has a large impact on everything else in the company, and so on.

There are many unknowns in the early days of building a company, and considering different models of success can be a useful exercise for forcing you to be more explicit about the weights you assign to different objectives and near-term tasks.

Thank you to Sean Byrnes and Tyler Hogge for giving me feedback on this post.

** Just for fun, here are three examples of these models’ weaknesses:

1. The first model assumes all risks are independent, even though they’re not (e.g. stronger product-market fit often implies a better product, and a better product reduces sales risk by being easier to sell).

2. In the second model, dividing market size by the number of competitors is unlikely to be correct. A $10b market with 10 competitors will result in one company owning most of the market and everyone else struggling to catch up; it’s not likely to result in 10 companies worth $1b each. For example, consider Uber vs. Lyft vs. everyone else.

3. In the third model, there are many milestones that can be achieved in parallel or that do not depend on each other. For example, you can build a sales team and explore marketing channels at the same time, and making progress on either goal will increase your company’s valuation.

What? 25-minute time slots for office hours with a software engineer turned VC.

When and Where? 10am-4pm on Dec 28 (Redwood City), Jan 4 (Redwood City), and Jan 11 (SF).

Who is this for? Current and aspiring founders in the SF Bay Area. (If this experiment is successful, I’m open to online office hours in the future.)

How to Apply? If you’re interested, please email me with 1-3 topics that you’d like to discuss and a link to your LinkedIn profile (so that I have a sense of whom I’m talking to).

(Anticipated) FAQ

What can we cover during office hours?

Some things I can offer during a session: feedback on your business/idea/pitch, general business advice, my impressions of where your company’s biggest risks are, opinions on what the market valuation for your next round might be, etc. I’ll honestly answer any question that you have, or I’ll tell you that I don’t know the answer.

What’s your background?

I studied computer science at Caltech, then spent ten years working as a software engineer. During that period, I was one of the first engineers at LinkedIn, and then spent three years at Google and four years at Factual. For the last four years, I’ve been a cofounder/partner at a seed-stage venture fund called Susa Ventures. We’ve invested in about 50 companies so far, including Andela, Flexport, LendUp, and Robinhood. Most of my investing experience is with seed-stage B2B companies.

My LinkedIn profile.

Why should I care about your opinion?

You probably shouldn’t. But hey, it’s free!

How will you pick who gets a time slot during office hours?

More likely: there are fewer people that want time slots than there are time slots. Everyone gets a slot.

Less likely: if there are more people than time slots, there will be a preference for discussion topics where I think I can be more helpful, and for people who come from groups that are underrepresented in the founder community (when emailing me, you can optionally let me know if you fall into such a group).

Can I use my office hours slot to pitch you for an investment?

No. If you want to practice your pitch and get feedback then I’m happy to help. If you want to do a formal pitch because you think my fund may be a good fit for your seed round** then office hours are not the right venue. Just email me about doing a formal pitch instead.

Can you refer me to other investors after we talk?

Sorry, no. Most VCs only want intros that are warm/vouched for, and I can’t offer intros like that if we’ve spent <30 minutes together.

What kind of things would you like to learn?

Almost anything: what it’s like to work in a field I’m unfamiliar with, your thoughts on the industry you work in, your favorite management advice, etc.

Will you reply to every email?

I’ll try! I have no idea how much interest there will be. In the likely scenario that I get 1 or 10 or 50 emails, I’ll reply to all of them. In the very unlikely scenario that I get 500 or 1,000 emails, I might not be able to reply to everyone, but I’ll do my best.

Will you make office hours a regular thing?

If the experiment works out well, then yes.

Footnotes

* Founders assume VCs don’t work during the holidays, so they don’t try setting up pitch meetings. As a result, VCs don’t have any work to do and end up going on vacation. =\

** Roughly speaking, Susa Ventures invests $500k-$750k into seed-stage companies. We mostly invest in B2B companies, like to see a little bit of traction/market validation (e.g. $10k+ in monthly revenue), and we limit our focus to companies that have the potential to build strong long-term competitive advantages (network effects, proprietary data, economies of scale, and so on).

“In preparing for battle, I have always found that plans are useless but planning is indispensable.”

- Dwight D. Eisenhower

In the early days of building a company, it’s hard to step back and think about the bigger picture when there are always so many fires to fight or opportunities to chase down. That’s a shame, because looking at your business from a 50,000-foot view often reveals areas of misplaced focus or resource misallocation.

One technique that can break through the bubble of always being in the weeds is conducting occasional thought experiments that give you insight into your company. This post is a catalog of sixteen useful thought experiments. Each experiment consists of a thought-provoking question and an explanation of that question’s purpose/motivation. Most of these thought experiments require a different frame of reference on your company – either from the perspective of another person or another time period. Many of the experiments directly correspond to the risk categories discussed in my previous post, How to De-Risk a Startup.

Caveat: some of these experiments are aimed at seed stage b2b startups, but most are applicable to the majority of startups.

List of Thought Experiments

Looking at the present

• How do competitors view you?

• Are you ready for a Series A?

• The other side of the table

• The debate from hell

Looking back from the near future

(One year from now.)

• Was your MVP truly minimal?

• Stomach-churning churn numbers

• The missing key

• Laughed out of the room

• Sunk Cost Fallacy

• Misplaced Focus

Looking back from the distant future

(5-10 years from now.)

• Unexpectedly large market

• Unexpectedly small market

• The world of the future

• The badass competitor

• The nightmare competitor

• The disgruntled competitor

Looking at the present

How do competitors view you?

Think about your 2-3 closest competitors. If you asked their CEOs to truthfully describe your company’s strengths and vulnerabilities, what would each of them say?

Motivation: Competitors will view you differently based on their traction, their leadership, and their own strengths and weaknesses. For the most part, your competitors are not dumb and they are not just going to lie around waiting for you to steamroll them. Looking at yourself from their perspectives can give you a better sense of your own company’s strengths and weaknesses.

Are you ready for a Series A?

Imagine you suddenly have access to an additional $10m of capital to scale your business. Would you know what to do with it? If not, what would you need to figure out before you could make good use of more capital?

Motivation: Most Series As provide capital to double down on what’s working. If you’ve figured out some good strategies for sales and marketing and need to staff up in those departments, or if your engineering team is struggling to keep up with paying customers’ feature requests, then those are good reasons to raise more money. If you need more capital to keep the company alive but don’t have a clear plan for using that money then you’re probably not ready for your next major round of financing.

The other side of the table

Imagine that instead of being the founder of your company, you’re an investor who doesn’t know anything about the company or the market it’s pursuing. What would you want to know to understand the business? What questions would you ask? Where would you be skeptical? If you invest in 1%-2% of the startups that you research, what would get you excited enough to invest here?

Motivation: It’s easy to forget that other people – including investors – know much less about your business than you do. Your pitch should be designed to get investors excited about the opportunity and to allay concerns they might have. Thinking about your company from a layperson investor’s perspective can help you understand which aspects of your company should be emphasized and which ones need longer explanations.

The debate from hell

You’re in a debate competition. The subject of the debate is the viability of your startup (whether the problem being addressed is legit, whether the market is big enough, whether your approach is right, et cetera). You’re taking the “pro” side and an award-winning debater is taking the “against” side. Your opponent does their best to rip your idea to shreds. What arguments does your opponent make? How do you respond?

Motivation: during first meetings, most investors are looking for “reasons to say no.” An investor might only invest in 1% of the companies they see, so a typical investor process is to pass on >80% of pitches quickly so that more time can be spent on the remaining 10-20%. Considering your startup from the point of view of someone who is (initially) trying to dismiss your idea can help improve your pitch and address its weaknesses.

Looking back from the near future

For the following thought experiments, pretend that you’re looking back on your company one year from now.

Was your MVP truly minimal?

You’ve hit product-market fit. One of your learnings is that your Minimum Viable Product ended up going far beyond “minimal.” Which features could you have left out?

Motivation: People often overestimate what’s required for an MVP. An MVP can be as simple as a founder armed with a cell phone and a spreadsheet. Stepping back and thinking about your product from a future perspective might give you ideas for what you can cut before launching. The sooner you launcher, the sooner you can learn about what your customers really want.

Stomach-churning churn numbers

Half of the customers that signed up last year decided not to renew their annual contracts this year. Why did that happen? What parts of your product/service/customer experience made them leave?

Motivation: Once they starting booking revenue, companies sometimes over-focus on growth while neglecting engagement and retention. Thinking about why users might not renew can help avoid churn issues before they become serious.

The missing key

You realize your company would’ve made a lot more progress if you had hired earlier for a key role. What was that role? What were the signs that you should’ve hired someone for that role a little sooner?

Motivation: Because time and money are so limited at most companies, founders will often take over jobs they’re unfamiliar with in order to keep costs manageable. That might mean the CEO is doing sales despite coming from engineering background, or it could mean the CTO is doing basic company accounting. Wearing many hats is a necessity at startups, but sometimes you need to hire a professional for a certain hat. Thinking about the “appropriateness” of what founders and employees are doing each month can uncover areas where additional full-time or part-time hires are necessary. There are often early warning signs that some function is being handled very inefficiently or even dangerously.

Laughed out of the room

You recently started your fundraising process. One of the most respected funds that you talked to (e.g. Sequoia, Benchmark, etc) said your pitch is one of the worst they’ve heard all year. What made them say that?

Motivation: The earlier “debate from hell” thought experiment was about looking for flaws in your business plan. This experiment is about finding flaws in your pitch, like claims that are true but which sound crazy or implausible, or ideas that require much more explanation than what you typically offer.

Sunk Cost Fallacy

As you look back on last year, you realize you spent too much time on several initiatives because of sunk costs. What were those areas, and what were the signs that you should’ve cut your losses instead of continuing to waste more resources?

Motivation: People are prone to continuing dead-end projects because they’ve already invested so much time and effort into them. This is especially dangerous at startups, which have very limited time to reach key milestones before running out of money. This thought experiment is designed to help you see if sunk costs are causing you to waste your time.

Misplaced focus

As you look back on last year, you see periods of time where your company focused on the wrong things. Reflecting on these periods, you realize that you overlooked many signs that your resources were being wasted. Where did you waste your time and what were the signs that you should’ve been working on something else?

Motivation: Sometimes you miss signs that things aren’t working. Maybe engineers keep going to Google or Twitter instead of accepting your job offers, or maybe very few customers agree to see your demo after an initial discovery call. When you’re scrambling day-to-day, you might think, “if 5% of people want to see a demo, then I should call at least 40 people daily.” Months later, you realize that the low demo rate was a sign that your product didn’t fit the market’s needs. It’s helpful to step back and ask yourself if the things that you’re struggling with today are a sign that you need to optimize or double down on your processes, or if they’re a sign of something more significant, like working on the wrong product or targeting the wrong job candidates.

Looking back from the distant future

For the following thought experiments, pretend that you’re looking back one your company 5-10 years from now.

Unexpectedly large market

You’re company is very successful at this point, and you’ve learned that your target market is actually 10x bigger than your original estimate. Where did the extra market size come from?

Motivation: This experiment might help you see opportunities in selling to adjacent target audiences, or selling additional products to your existing target audience.

Unexpectedly small market

Your company’s growth has plateaued, and you’ve learned that your target market is actually 10x samller than your original estimate. Why did the market initially seem so much bigger than it actually was?

Motivation: This is a good way to scrutinize how many people have the problem you’re solving, and whether the problem is severe or mild. It’s tempting to throw out impressive market size statistics, but sometimes numbers are misleading and you waste years of your life chasing a market that turns out to be unattractive.

The world of the future

How does the trajectory of the world over the next 5-10 years align (or misalign) with what you’re doing? Which trends will help you and which trends will hurt you? What’s the expected timeline for those trends?”

Motivation: an idea that is good or bad today might look very different 5-10 years from now. For example, solving parking problems is a big deal in 2016, but perhaps fleets of autonomous vehicles will eventually make parking garages obsolete. Thinking about relevant trends can help you pursue the optimal path for your company. It would be a shame if you found a solution to a major problem and started scaling your company just as technological or regulatory or social trends made that problem disappear.

The badass competitor

Elon Musk, no longer content with developing solar-powered autonomous vehicles that roam the surface of Mars, decides that his next company will be a direct competitor to you. What barriers to entry and competitive moats have you built up that will keep Elon’s efforts at bay?

Motivation: A lot of founders think their special sauce is how well they execute. The problem is that almost everyone thinks they have superior execution skills. Thinking about what you would against competition that executes even better than you can provide clarity about your true strengths, as well as how you can apply those strengths to build a strong competitive moat around your company.

The nightmare competitor

After you raised your $100m Series E, a lot competitors emerged. Everyone wants a piece of your pie. Most of these competitors don’t worry you, except for Company X. Company X has tons of cash and is in a particularly strong position to compete with you. What makes them so formidable?

Motivation: There are often one or two companies that would be particularly intimidating as competitors. These companies might have great access to your target customers, tons of world-class talent, or a product that lets them get between you and your users and slowly take over your relationship (e.g. how Google uses its search engine to highlight its own reviews at the expense of Yelp and TripAdvisor). It’s important to be mindful of the companies that are best positioned to compete with you. These companies might not care about your market today, but they could change their minds in the future once your company has reached an enviable size. Knowing which companies to be cautious of can help you plan and set up barriers that prevent those companies from pursuing your market.

The disgruntled competitor

After almost a decade of working together, you just had a falling out with one of your key execs. They quit their job and decide to start a company that will put yours out of business. What actions have you taken over the years to address this threat?

Motivation: This is subtly different from the “fearsome competitor” scenario. That scenario is focused on realistic worst-case competitors. This scenario is focused on the hypothetical worst-case competitor: someone with funding, a chip on your shoulder, and insider knowledge about everything related to your company.

How to use these thought experiments

While not all of these questions will apply to every company, it’s a good exercise to think about them one-by-one. By analyzing your company from different perspectives and asking yourself challenging, through-provoking questions, you’ll be able to have a better understanding of your strengths and weaknesses, areas where you might be mismanaging your resources, and future opportunities.

The best way to get value out of these thought experiments is to occasionally pick a question and think about it for 5 or 10 or 30 minutes. If a train of thought proves useful, discuss it with your coworkers and revisit the question every few months.

I hope you will find these thought experiments to be a helpful exercise for moving your company forward. If there are other thought experiments that have been valuable for you, please let me know on Twitter.

This post contains a (non-exhaustive) list of common startup-related risks, the spectra along which those risks might be classified, and some tips and heuristics for mitigation. The further you move from “high risk” to “low risk” along each spectrum, the stronger your valuation, perceived progress, and likelihood of success will become.

The entries on each risk spectrum are rated from 1 (high risk) to 5 (low risk). Your goal is to move away from the 1’s and toward the 5’s. For example, if you were trying to prove to a friend that you could bake a great cake, then the risk spectrum might look like this:

• [1] You’ve never baked a cake before, but you’re sure you could do a good job.

• [2] You show your friend a picture of a nice cake you made in the past.

• [4] You introduce your friend to people who have previously tasted your cakes and loved them.

• [5] You let your friend taste an amazing cake that you just made.

High-level Principles

Most of the example risk spectra that follow can be summarized by three core principles:

Principle #1: showing is better than telling.

• [1] You think you can do XYZ.

• [3] You’ve done XYZ in the past.

• [5] You’re currently doing XYZ, and doing it well.

Principle #2: external validation is stronger than your personal opinion.

• [1] You claim XYZ.

• [3] Numerous people who are affiliated with you (friends/accelerator batchmates/etc.) claim XYZ.

• [5] Numerous people who are completely unaffiliated with you claim XYZ.

Principle #3: more data is better.

• [1] Your product has 0 sales.

• [3] Your product has 5 sales.

• [5] Your product has 50 sales.

9 Risk Spectrum Examples

As a caveat, the following are based on my personal risk “ratings.” Other people might look for different proof points or have different ratings for the same proof points.

1. Product/Market Fit Risk

Goal: prove that you’re actually building something that people want.

B2B version:

• [1] You think people will want to use your product.

• [2] You talked to potential customers, and they said they wanted to try the product once it was built.

• [2] You have LOIs (letters of intent).

• [3] You have unpaid pilots.

• [4] You have paid pilots.

• [5] You have paid contracts. Ideally prepaid.

B2C version:

• [1] You think people will want to use your product.

• [2] You have some early users, but they’re all affiliated with you (friends, family members, etc).

• [3] You have some early unaffiliated users, but user acquisition economics aren’t great.

• [4] Your user base is growing organically at a moderate rate.

• [4] Your user base is growing quickly through affordable paid acquisition.

• [5] Your user base is exploding through referrals and word of mouth.

Another perspective:

• [1] Engagement is very low, churn is very high.

• [3] Engagement and churn are both moderate.

• [5] Engagement is very high, churn is very low.

2. Product Quality Risk

Goal: prove that you can build a great, high-quality product.

• [1] You are sure you can build a great product, but haven’t built one before.

• [1] You plan to outsource product development.

• [1] You have a prototype, and it’s very mediocre.

• [2] You’ve previously been on a team that built great products.

• [3] You previously led a team that built a great product.

• [4] You have a prototype, and it’s good.

• [5] You have a live, fully-functioning product and it’s amazing.

3. Team Risk

Goal: prove that you’ve assembled a great team for achieving your vision.

Your product requires strong execution across many functional areas (eng, sales, UX design, etc.) and…

• [1] …your full-time team only covers 1-2 of those areas.

• [3] …your full-time team only covers 1-2 of those areas, but investors and advisors fill in most of the gaps.

• [4] …your full-time team covers most of those areas.

• [5] …your full-time team covers all of those areas.

4. Recruiting Risk

Goal: prove that you will be able to grow your team effectively. (This is a very real risk in Silicon Valley, where demand for good engineers is much higher than supply.)

• [1] You’ve never hired anyone before.

• [1] You’re not someone people would want to work with. (E.g. you’re a jerk.)

• [2] You have some interviewing experience.

• [3] You have prior recruiting and management experience.

• [4] You’ve built very strong teams in the past.

• [4] Your team already includes several great hires.

• [5] You’re currently able to reliably recruit in-demand candidates through personal charisma, a strong company mission, an amazing company culture, or something similar.

5. Sales Risk

Goal: prove that your team can sell the product effectively.

• [1] No one on your team has any sales experience.

• [2] You’ve done sales, but not much or not recently or not of the same flavor that you’ll need for your company.

• [3] You’re successfully selling you product, but it’s significantly underpriced and/or sales are taking much longer than expected.

• [3] You’ve done a lot of sales work that’s similar to what you’ll be doing at your startup.

• [3] You’ve built and led successful sales teams in the past.

• [4] You have a strong, experienced sales team.

• [5] You’re successfully selling your product at a good price and with reasonable sales cycles.

6. Market Risk

Goal: prove that if you execute well, you can make enough money to become a huge company (e.g. $1b+ exit potential).

• [1] Your target market is small and unlikely to grow quickly.

• [1] You have no idea how large your target market might be.

• [3] You found a Gartner report that gives an estimate of market size.

• [4] You have a plausible top-down analysis of market size (“People spend $X per year on this problem, and we think we can capture 15% of that with our solution.”)

• [4] You have a plausible bottom-up analysis of market size (“We think we can capture 10% of users in group A and 20% of users in group B, and we plan to charge those users $X and $Y, respectively.”)

• [5] You have a plausible bottom-up analysis, backed by experiments and data. (“There are X million potential users for our product, and we’ve done some tests that show each user would be willing to pay $Y/month”)

• [5] Incumbent companies are huge and demonstrate that there’s a big market for what you’re doing.

7. Funding Risk

Goals: prove that you have enough capital to reach the milestones needed to raise more money on better terms (if you want to), and that you have a viable back-up plan if you can’t raise money on your ideal timeline.

• [1] Your business will not be self-sustaining for a long time, and you’re completely dependent on raising many rounds of venture funding.

• [1] You have multiple well-funded competitors and will need tons of capital to compete with them.

• [2] There’s no cushion between when you expect to have enough momentum to raise your next round and when your current cash balance will force you to try raising another round. (E.g. you think you’ll need 12 months to get to $1m in revenue and 3 months to raise a Series A, and you have exactly 15 months of runway in the bank.)

• [2] You’ve successfully raised some venture capital before.

• [3] You’ve successfully raised tons of venture capital before.

• [3] You have good funds investing in your current round, and they will help guide you to your next round.

• [4] You have good funds with deep pockets investing in your current round. (That is, funds that could lead your future rounds if they wanted to.)

• [4] With some effort and sacrifices, you would be able to get to break-even without any additional capital.

• [5] You’re not dependent on additional capital because you can easily become break-even or profitable at any time.

8. Short-Term Competition Risk

Goal: prove that you’re differentiated from existing players in the market.

• [1] There are many competitors of all sizes (huge incumbents, young startups, well-funded startups, etc). These companies are attacking your target market from many directions.

• [2] There are many competitors, but they are ineffective legacy players or poorly-funded startups.

• [3] There are very few competitors, but no strong differentiation between you and them.

• [4] There are very few competitors, and strong differentiation between you and them.

• [5] There are no competitors and there’s a high barrier to entry (which you have crossed).

9. Long-Term Competition Risk

Goal: prove that as you become successful and other companies try to copy you, you will be able to maintain your strong position.

• [1] You’re not the first mover and you don’t have a real competitive advantage.

• [2] You don’t have a real competitive advantage, but at least you’re the first mover.

• [3] You have weak competitive advantages: a few small patents, slightly better unit economics than new entrants, etc.

• [4] You have moderate competitive advantages: good brand perception among your customers, significantly better unit economics, a strong patent portfolio, etc.

• [5] You have strong competitive advantages, like network effects or proprietary datasets, that get stronger as you grow.

How to De-Risk

Step 1: Do an honest self-assessment of your company’s major risk areas.

Step 2: List ways to move from high risk to low risk along each risk spectrum. If you’re not sure what you can do, ask investors, advisors, or other founders.

Step 3: Create short-term and long-term risk mitigation plans for your company. Some of the action items might be small and easy, like running an AdWords test to estimate customer demand and market size. Other actions will take time, like figuring out an elegant way to bake network effects into your product or finding a good VP of Sales when no one on your team has sales experience.

Miscellaneous Tips

• If there’s little remaining risk in an area, focus on other areas. It is usually better to turn a few 1’s into 3’s than it is to turn a 4.5 into a 5.

• Good investors and advisors are an effective way of moving from a 1 to a 2 or 3. Maybe you don’t know how to hire a VP of Marketing, but if one of your investors or advisors does then at least that’s something.

  Corollary #1: a great complementary founder can take you from a 1 straight to a 4 or 5 in some areas. That’s a great way to boost your seed round valuation.

  Corollary #2: pick your investors and advisors deliberately. People who can help in areas where you’re lacking are more valuable than people with the same strengths as you.

• Get external feedback. Self-awareness is hard. Soliciting risk assessments from people you trust who will be honest with you is a great exercise.

• Try to address risks before you have to face them head-on. If you’ve never done sales, you can practice as soon as your product is ready. But you can also start practicing pre-launch, or even on a hobby project before you start your company in the first place.

Ultimately, addressing risks is not something you should do for investors, it’s something you should do for yourself. If you’re thinking of dedicating years or even decades of your life to something, it’s worth understanding where your biggest challenges will be and how your can incrementally address those challenges.

“In the South Seas there is a Cargo Cult of people. During the war they saw airplanes land with lots of good materials, and they want the same thing to happen now. So they’ve arranged to make things like runways, to put fires along the sides of the runways, to make a wooden hut for a man to sit in, with two wooden pieces on his head like headphones and bars of bamboo sticking out like antennas—he’s the controller—and they wait for the airplanes to land. They’re doing everything right. The form is perfect. It looks exactly the way it looked before. But it doesn’t work. No airplanes land. So I call these things Cargo Cult Science, because they follow all the apparent precepts and forms of scientific investigation, but they’re missing something essential, because the planes don’t land.”

- Richard Feynman, 1974 Caltech Commencement Address

The passage above may as well be describing the world of startups. While Silicon Valley is a magical place that brims with ambition and innovation, it’s also full of Cargo Cult behaviors. Famous founders and investors are idolized and mimicked by their colleagues, who put on wooden headphones and are surprised when that doesn’t lead to success.

Cargo Cult Segway by Peter Moosgaard

Cargo Cult Causes

Everyone in the startup ecosystem, from investors to founders to employees, is prone to Cargo Cult thinking. People try to extract lessons from the successes of others, but fall prey to numerous biases and fallacies:

• Survival bias – attributing causality to common attributes of successes without determining if those attributes were different among failures. Example: “The most successful CEOs are incredibly hardworking.” (It turns out that many unsuccessful CEOs are incredibly hardworking, too, so working 100-hour weeks may not be a good predictor of success.)

• Recency bias – overweighting recent events. Example: “E-commerce is an amazing sector to invest in. Just check out Jet.com and Dollar Shave Club!” (Three months ago: “E-commerce is an awful sector to invest in. Just check out One Kings Lane and Fab.com”.”)

• Confirmation bias – only noticing data points that support preexisting beliefs. Example: “FB ads are a great channel. Just last week we were able to acquire 3 great customers through FB.” (But what about the 9,997 potential customers who didn’t click on your FB ad because they mostly hang out on LinkedIn? Or what if it turns out you could’ve acquired the same 3 customers for 1/2 the cost by going to an industry conference?)

• Observational selection bias – noticing a lot of examples of something once we become aware of it. Example: “Sometimes there’s so much work that I end up sleeping at the office. Last week, two of my friends couldn’t meet up with me because they had to stay overnight at work. Working overnight must be a really common behavior!” (But what about your 500 other friends who didn’t work overnight last week?)

… and so on …

Our brains are wired to take examples of success and extract “lessons” from them, but these lessons often range between innocuously wrong and dangerously wrong.

The remainder of this post covers Cargo Cult examples, warning signs, and potential remedies.

Cargo Cult Examples

The common thread in all of the examples below is not that these behaviors are always wrong. (They’re not.) It’s that these behaviors are often wrong because they’re blindly copied from others without critical examination. There are many great reasons to use MongoDB or hire 10x engineers or start a VC blog, but “because that’s what everyone else is doing” is not one of them.

Investor Cargo Cult examples:

• Investing time into content marketing or social media. If you have something interesting to say, great! But don’t start blogging just because your peers did.

• Demanding a specific ownership level for each investment. A lot of VCs have specific ownership targets, like 10% or 20%. Why 20%? Why not 21% or 18%? If you try to buy 20% of Pinterest but can only get 18%, is it really better to walk away?

• Having a policy of requiring warm intros and not replying to cold emails. This is a reasonable system for very busy investors, but many investors are not that busy. And some cold emails are pretty damn good.

• Co-investing blindly whenever a top VC is investing. While top VC firms have a lot of hits, they have a lot of misses, too. Trying to co-invest with them on every deal is a mistake.

• Trying to be contrarian. Being contrarian is overrated. A lot of the most successful companies, like Google and Airbnb, looked like crazy investments at first. But other great companies looked very sane and likely to succeed, even in their early days. An investor can accumulate a great portfolio without being contrarian.

• Being intolerant of tech outsourcing and only investing in companies with technical cofounders. Companies like Slack and Github outsourced pieces of their initial products. Other companies, including Pinterest and Whisper, didn’t have engineering cofounders. Depending on the product, a startup that uses outsourcing thoughtfully can become successful without having engineers on the founding team. (Hat tip to Phil Martie for this Cargo Cult example.)

Founder Cargo Cult examples:

• Joining accelerators. YC is awesome, but you don’t have to go through YC to build a good company. I’ve actually seen companies that have gone through 3-4 consecutive accelerator programs. (Spoiler: that will not make your company 3x-4x more likely to succeed.) Some accelerators are great, but others take equity and don’t provide any true value. Don’t go through a program just to have a logo for your pitch deck or because that’s what you think you’re supposed to do.

• Raising money. Raising money is a strategic tool, not a prerequisite for building a successful company. Too many founders assume that they can’t start working toward their vision until they’ve raised outside capital.

• Optimizing for valuation. Raising at too high (or too low) of a valuation can have severe downsides, but founders often try to match or beat the valuations that they’ve seen their peers receive.

• Limiting hiring to 10x engineers. Why 10x? Because that’s who everyone else is trying to hire. The truth is that a lot of companies can be built with 4x or 2x engineers, and sometimes even 1x engineers.

• Offering trendy employee benefits. Kind bars. An office in South Park. Yoga classes. Martini Mondays. If you explicitly want to build a culture around these things that’s fine, but you don’t have to offer these perks just because other startups do. Plenty of employees love doing meaningful work even if it doesn’t come with fancy granola bars.

• Getting advisors. Many startup pitch decks list advisors who are either unrelated to a startup’s business, or are high profile people that are so busy that they’re not going to be involved. Not having advisors is better than having advisors just for the sake of having an Advisory Board slide.

• Only reaching out to investors via intros. This is a second order Cargo Cult behavior. Because investors strongly prefer warm intros, some founders assume that any intro is better than a cold email. Not true. An intro has to be warm. That is, the introducer should know both the founder and the investor well. Otherwise, the intro is worthless because either the introducer can’t truly vouch for the founder, or they can vouch for the founder but their testimonial carries no weight with the investor.

• Copying the strategies of successful competitors. Every startup has to write its own playbook. Content marketing or speaking at conferences or recruiting at meetups might work for another company, but that doesn’t mean those things will work for you. Blindly copying other companies usually leads to failure.

• Working harder when things aren’t going well. People work 80- or 100-hour weeks to save a dying company because they think that’s what they’re expected to do. Sometimes the problem is in what you’re working on or in your approach, not in how hard you work.

• Mimicking big company processes, like code reviews or performance reviews. A process that works well at Salesforce or Facebook is often inappropriate or dangerous when implemented at a startup. Having formal performance reviews at a 5-person company is just as bad as not having them at a 5,000-person company.

Cargo Cult Warning Signs

• People inquire about why you’re doing something, and your immediate response is 1) it’s what everyone else does, 2) it’s what your role model does, or 3) it’s what you think is expected of you.

• You’re afraid to start/stop doing something because that would be different from what most people do.

• You invest a lot of your time and resources into an activity, but you’ve never carefully considered whether the effort is worthwhile.

Tips for Diagnosing and Stopping Cargo Cult Behaviors

Cargo cult thinking is fiendishly hard to diagnose. Imagine you’re the tribe describe in Richard Feynman’s example: how would you figure out that building wooden headphones is not a good use of time? Here are some tips and ideas:

• Question your assumptions – especially for large decisions. Why did you decide to spend your entire marketing budget on Facebook ads? Why do you shy away from investing in solo founders? Why do you want to work at a seed stage startup? If you have personal reasons for doing these things, that’s great. If your reasons are based on what you’ve seen others doing, that’s a red flag.

• Get feedback from a diverse group of unbiased observers. It might be hard to realize that wearing wooden headphones has no effect, but if you ask enough people for feedback, eventually someone will point out that you’re wasting your time.

• Actively seek out opinions of people who are willing to be honest and critical to your face. It’s very useful to have a few people you can turn to who like to play Devil’s advocate and question everything. Devil’s advocates are a vaccine against Cargo Cult behaviors. These people could be employees, friends, or even members of online communities like HackerNews and Twitter.

• Listen to critical questions. Sometimes people around you will question you about why you’re doing what you’re doing. It’s tempting to get defensive or to give a canned response, but if you repeatedly hear the same questions, that’s a cue to become introspective about your behavior.

Once you identify something as a potential Cargo Cult behavior, analyze your assumptions and reasoning in depth to decide whether you need to make a change.

Finally, here’s a useful exercise: list out 10-20 areas where you spend the most time, money, or effort. For each item on the list, determine why you allocate so many resources to it, and whether your reasons are well-founded or if you’re just acting on autopilot.

TL;DR: just because “everyone is doing it” doesn’t mean you should.

If you have Cargo Cult examples or tips that you’d like to share, please let me know on Twitter and I will update this post accordingly.