The three moats of a defensible, scalable internet startup

Aim for data over features

It’s not that complicated to build a small lifestyle business. Find an existing idea that works — any small business that generates profit — and create something similar. Market yourself a little bit, capture a part of the market, and you’re literally in business. Like many people, I’ve done it a couple of times in my career.

Internet startups are different. They’re not typically designed to be small businesses. They’re designed to take advantage of relatively frequent and early rounds of capital to accelerate growth and become large, $100 million+ enterprises.

Like the movie and music industries, the internet startup world is hit-based. An investor doesn’t invest $100,000 on the chance that they might get $150,000 or $0 back. They invest on the chance that they might get $1,000,000 or $0 back. That’s a risk worth taking. They want home runs, so the investment infrastructure that exists for startups is specially crafted to double-down on potential hits and let everything else die.

Due to these differences, you shouldn’t build an internet startup like you would a small business. Your startup needs to capture, and maybe even generate a large part of the market, keep it, and then keep going even faster. It needs to be defensible and almost infinitely scalable.


Features are an easy way to build a small business: they’re usually straightforward to copy, and simply marketed to an audience that understands their benefits. But it’s difficult to scale based on features alone, even with more or better features. You have a limited amount of the market’s attention to try to describe your new features, and why they’re necessary. That’s a lot more difficult than it sounds. You’re also still going to be compared to your competitors, questioned on your price, and you open yourself up to the same easy facsimiles.

What’s more, as a startup, your established competitors almost certainly have more experience, resources and market share than you do to benefit from any valuable new features that you identify for them.

You’ll win some customers and lose some customers, and may even grow at a reasonable pace, but features alone aren’t likely to give you the massively scalable, defensible product you need to grow exponentially.

Design, Technology and Data. But mostly Data.

A look at the spectrum of successful, high-growth internet startups seems to suggest that there are three main strategies that can give you scale:

Pyramid of scalable internet startup strategies

Design – not necessarily simple impulsive desirability or aesthetics, but often a radical take on user experience that makes a traditionally complex system simple to use and understand. Examples include MailChimp, GitHub, RelateIQ, Mint and Tinder.

Technology – not a single feature, but the development of a difficult, invisible “magic” that powers the startup and cannot be easily copied. It’s an invention. Examples include Google, Skype, Palantir and Kaggle.

Data – can be an aggregation and centralization of previously disparate or offline data sources, or it can be valuable proprietary data that has no or limited availability elsewhere. There are many examples of this type of company, including eBay, Airbnb, Facebook, LinkedIn, Shutterstock, Netflix and Yelp.

Some companies span two or three of these strategies. Google, for example, originally designed a strikingly minimal interface to a powerful underlying ranking technology. They have since built an unimaginably large data store which gives them an untouchable understanding of the connected web of information, with which they can further improve their technology.

In fact, this seems to frequently be the case. Startups will initially use superior design to persuade the market to trust them with their data, or develop a technology that expedites data collection. It suggests that data is the ultimate defensible startup strategy.

Some people may say that “information wants to be free” but if we look at the modern movement of data, it seems that it actually prefers to clump together with similar data in dark, secret silos. To use a pseudo-mathematical metaphor, it’s almost like minima in an energy plane, where data of similar types inevitably rolls down hill and settles at the place where it is most easily used.

Thinking about this at the largest scale of business – public companies – it means that I’d much prefer to invest money in data-heavy Facebook, LinkedIn or Google than data-weak Groupon or Zynga.

Change in share price since Facebook IPO of some major internet stocks

Stock price change since Facebook IPO date

Groupon may have had an explosive start when they had an edge on design, but they have so far been unable to convert their high volume of usage into a meaningful proprietary data store. Thanks partly to Groupon’s initial success, dozens of middle-men businesses sprung up that sourced and sold coupon data for small businesses, allowing numerous Groupon facsimiles to be built almost overnight.

And while Zynga may occasionally have one or two hits through superior design, their success doesn’t appear to be contributing to a growing data source that would give them a lasting competitive advantage. Video game companies are notoriously difficult to grow over a protracted period of time. As a form of hit-based entertainment, it would be like asking a rock band to grow in popularity forever.

So, if you’re building an internet startup, ask yourself which scalable strategy or strategies you are using, and if data isn’t one of them, how you might evolve so that it eventually is.