Invisible builders

Inaction is the real competition for infra companies

We’ve been lucky to support builders alongside dozens of amazing infrastructure companies over the past year of running Build, and one thing is true across the board — they see massive dropoff of builders from awareness to adoption. 

But interestingly, companies typically ascribe builder dropoff to poor documentation and/or an uncompetitive product. They assume they lose builders to better solutions.

There is a far simpler answer that describes why all infra companies that grow bottoms-up have leaky funnels — most builders simply give up. And the ways things are going, that’s only going to get worse.

Learning from MOOCs

Massive open online courses (MOOCs) were all the rage for a while, until studies came out showing that completion rates are typically somewhere around 3-5%. This isn’t a specific phenomena of MOOCs though. It’s just an uncommonly clean dataset on online, unstructured building in general. 

The same pattern exists everywhere in self-motivated projects. 90% of podcasts don’t publish more than 3 episodes. 92% of people don’t complete their new year's resolutions. There isn’t great data on the completion rate of launching a side project or startup, but a safe bet is that the rate from idea -> product-with-real-users is even worse.

People who abandon a course on building an app in React don’t typically find and complete a better course. Podcasters who stop after a few episodes don’t start shopping around for recording tools with better docs. They mostly just give up, either for good or with a significant refractory period before they try again. Much has been said about the trough of disillusionment for builders but not enough about it’s unfortunate sibling, the hangover of abandonment — after you unsuccessfully try to build something, it takes a while to rebuild that energy. 

The point is, for most builders the tech isn’t the biggest problem. So if you’re in the business of helping people build stuff — whether a product, podcast, habit, anything — you’re not just in a feature or documentation war with your competitors. You’re battling the simple fact that building something new is fucking hard.

You’re more like Stripe than you think

Part of the reason companies prefer to focus on features and documentation is that they are easier problems to understand and measure. Builders who abandon their project are invisible. They don’t send in feedback or bugs, they just show up as churn. So paradoxically, the more you “fix” the abandonment problem, the more it will feel like the real problem is better features and documentation (because you have more active builders). 

Despite the challenge with tracking progress, I believe great infra companies will increasingly treat abandonment as a key problem to solve, and in that sense more companies will be like Stripe. Stripe famously spends a lot of time and money helping builders get over the hump of launching a startup — inspiration from Stripe press, easier incorporation from Stripe Atlas, countless sponsorships of hackathons and other programs, and so on.

Stripe doesn’t just do these things altruistically, as many people seem to think. Stripe realizes better than most that if your business relies on bottoms-up growth of new products, your biggest competitor is inaction. And that will only become more true in the future — the completion rate of builders is about to get a lot worse. 

TikTok has made everyone a potential creator, it has also made everyone a potential abandoner of a creative project. The same is true of software. The rise of AI-assisted developer tools and the continued democratization of product building is amazing, but it is going to send conversion rates through the floor.

To be clear, nothing will ever replace the need to build a killer product as an infra company. Stripe is obviously a great example of that as well. But it’s not the only problem, and companies that fight the pull towards inaction instead of just fighting a feature war will increasingly have a big advantage.

The Maven-ification of Infra

Few companies are at the scale or market share where a Stripe Press type solution makes sense. A tide that lifts all boats works great if you have the biggest boat. So where will most infra companies experiment? And what new businesses might be created in that process?

To look for leverage, we can oversimplify what builders need to create something valuable: 

1. The right idea, that actually solves a problem that is worth solving

2. The right constraints, motivation, and support system

3. The tech/money/network to do it

That third bucket gets orders of magnitude more time and resources than the former two, and the first one around finding the right idea is a thorny problem that tech will never really be able to “solve.” So I would expect most of the low hanging fruit is in the second bucket — let’s call it creating the right containers for builders to be successful.

The most transformative containers we’ve seen so far are cohort type programs like Y Combinator. You can also bundle EIR programs, company startup programs, hackathons, and of course what we do with Build in there. How might infra companies borrow insights from those models to create containers for builders around their tech?

One place to look for inspiration is Maven, which is essentially Coursera but with a synchronous cohort model added on top. Maven didn’t achieve >15X better completion rates because of better tech or educational resources than Coursera, it simply turned online courses into synchronous cohorts. 

So what would the Maven-ification of infra look like at scale? And how will the opportunity space change as an increasing number of infra decisions are made directly or assisted by AI? What existing support systems from infra companies will crumble or become untenable as the number of people capable of building their ideas increases exponentially? I can’t claim to have all the answers, but those feel like the right questions.

As an aside, while I do think trying to give builders the right idea is generally a tar pit, there’s a lot of room for improvement to “Request for projects” which is how companies generally address that problem now. The issue we see is that companies typically prescribe solution directions, but what builders actually need is a clearer sense of good problems to solve. If I were building a list of RFPs for a company I would skip the solutions entirely and just create a list of personas and problems from real people that I think my tech might help solve. If you connect an engineer to a clear problem, the rest happens pretty naturally. RFS 100 and other similar experiments were in the right direction, although I still think too focused on the product rather than the problem/persona.

Long live the long tail

One of the common reasons I hear to not work on builder abandonment is that it’s an adverse selection problem. In other words, the best builders will make it through without any extra “support.” Low completion rate is not a “problem” in that view, it’s simply a reflection of the fact that most people are bad/stupid/lazy. 

I hope it’s obvious how untrue that is. For one thing, see the previously mentioned massive economy of systems that support builders — accelerators, EIR programs, founder communities, etc.. If you think the number of great companies that exist only because of those systems is zero, I can’t help you.

Notably though, most of those systems work via funding + equity. They are a direct extension of the traditional Silicon Valley venture model. But as the long tail of builders explodes, a smaller % of projects will actually make sense in that world. Infra companies are my bet for where a lot of the support for that long-tail of builders will come from, because they have the resources and incentive to do it. But it will require adapting those support systems into a new frame that doesn’t just run on equity.

One of our Build sponsors from day one, Coinbase, has created billions in value over the past year via their L2 blockchain, Base, not by creating a massively differentiated product right out the gate but by basically capturing developer mind share by — for lack of a better phrase — “out vibing” the competition and motivating a massive amount of builders to create something new. It’s impossible to know for sure, but many of those projects would likely not exist at all without the energy Base has created, despite the fact that many similar solutions to Base exist.

Put simply, great infrastructure companies don’t just fulfill existing demand from projects. They realize that the energy to build is not a binary inherent to a builder, but something that can be stoked and leveraged as a competitive advantage. They turn invisible builders into visible ones.

I am of course talking my book here — I have been thinking, writing, and experimenting with systems to support builders for many years. We founded Backdrop to do that, and while we also feel the pressure to continue building a killer product, I am keenly aware that truly unlocking innovation is a very human problem. I am convinced that our real opportunity, like Stripe and any organization that serves builders at early stages, is to fight abandonment.

If you are exploring or experimenting with this topic, especially given changes to how people build in the age of AI, I’d love to chat — please reach out.

Thanks to David Sherry for the help workshopping!