It’s important for companies to have some kind of durable competitive advantage – otherwise competitors can come in, copy a product, and take away significant market share (and profits).
There are many types of competitive advantages:
A strong brand (Coca-Cola, Apple)
Strong IP (IBM, pharma companies)
Network effects (Facebook, WhatsApp)
Economies of scale (Walmart, Intel)
Data assets (Amazon’s purchase and browsing data, Netflix’s view and rating data)
Accelerating and Decelerating Advantages
All of these advantages can be strong and can help a company persevere against both current and future competitors. However, one aspect that’s often overlooked is how these advantages evolve as a company grows. In some cases, these advantages accelerate as a company gains market share, in other cases they decelerate.
Accelerating advantages are great for incumbents because when a new entrant gains some market share, they only receive a tiny sliver of the benefits of scale. A huge market share is required to attain most of these benefits.
Decelerating advantages are great for new entrants because most of the benefits of scale can be captured with only a little bit of market share.
Here’s a graph that illustrates accelerating and decelerating competitive advantages:
Some concrete examples:
Manufacturing costs are a decelerating advantage. Your costs go down quickly as you move from making 5 units to 100 units. As you move from 100 units to 5,000 units, and then to 200,000 units, your costs continue to drop, but less and less rapidly.
Product catalog breadth is a decelerating advantage. As your store’s catalog expands from the top 100 best-selling products (for some shopping category) to the top 3,000 items, your shop becomes more appealing. However, going from 3,000 items to 10,000 has a much less dramatic effect, and increasing your offering further has less and less impact.
Network effects for communications tools like WhatsApp are an accelerating advantage. If 25% of your friends have WhatsApp, then that’s kind of useful. If 50% do, that’s a lot more useful – chances are if you need to reach a group of friends, at least one person in the group has WhatsApp installed. If 100% of your friends WhatsApp, then that’s terrific, as your can now talk to anyone you want with a single application. The competitive advantage is accelerating because 100% coverage is more than twice as useful as 50% coverage, which is more than twice as useful as 25% coverage.
A lot of competitive advantages are not solely accelerating or decelerating. For example, the usefulness of Yelp’s review volume accelerates at first, then starts to decelerate at a certain point. The more places have reviews, the more useful Yelp becomes. For example, if Yelp has reviews for 100% of all restaurants while Zagat has reviews for 50% of all restaurants, then Yelp will be many times more useful than Zagat. Furthermore, having an average of 50 reviews about each place instead of an average of 5 reviews is huge. However, at some point the value of more reviews starts diminishing. Having 500 reviews per place instead of 50 is a small improvement, and going from 500 to 5,000 reviews is barely noticeable.
Which Type of Advantage is Better?
The ideal structure of a competitive advantage depends on the stage of a company. If you’re just starting out, you want the competitive advantages in your target market to be decelerating. That is, you want to get most of the benefits of scale early on, so that you can compete with established players. This is how Quip and Google Docs managed to win market share from Microsoft Office. Although those tools had a tiny fraction of Office’s features when they started out, it turns out that having 50x as many features makes Office 2x better – not 20x or 100x better. The lower price and collaborative features of Quip and Google Docs more than made up for that 2x to a lot of users.
As your company becomes large, you want your competitive advantages to be accelerating, so that most of the benefits are realized only at a large scale. This creates huge barriers to entry and helps you maintain your market position.
So that’s the challenge: decelerating advantages make entering a market easy for you – but also for everyone else. Accelerating advantages make it easy to keep your lead in a market, but make it hard to replace an existing leader in the first place.
How to Shift from Decelerating to Accelerating
How can this dilemma be resolved? One approach is to try to ride a decelerating advantage as you enter a market, then add products and features to your offering to convert the advantage into an accelerating one. For example, you might start with a small, specialized product offering for your online store because that’s good enough for your initial customers. Over time, you’ll build logistics and manufacturing networks that make your costs significantly lower – but the returns from that diminish rapidly. So, you can start building recommendations and other features on top of your historical data that are impossible to replicate for someone without that data – now you’re turning your decelerating advantages (manufacturing costs + product catalog) into an accelerating one (great recommendations which require lots of data).
Another approach is to try avoiding a specific competitive advantage (e.g. the advantage of manufacturing costs) by introducing orthogonal features. For example, maybe you know you can’t compete on the cost of manufacturing cell phones with someone like LG Electronics, so you focus on software UX or hardware design or customer service or privacy features or something else – things that customers would not mind paying more for.
There are many other approaches companies can take to battle bigger players and suppress smaller players. If you have interesting ideas or war stories, please let me know on Twitter! And if you’re building a company, I believe it’s worth thinking about the areas of your market where it’s easier to get advantages of scale, and areas where you could block new entrants in the future because you have sufficient scale and they do not.
Further reading: Chris Dixon’s excellent post, “Come for the tool, stay for the network”