Please stop trying to build another Y Combinator

There are lots of startup accelerators out there. I believe that most of them add value to the startups they work with, as measured against the equity and time they take. In spite of that, most accelerators will not last long-term and could be adding much more value. Here I want to support development of alternative models, including non-profit and corporate-backed startup accelerator models, which I believe is where some interesting growth will be.

Blindly copying a model can be a bad idea

Most startup accelerators follow a similar model: a centralized program, taking somewhere between 2% – 10% equity for around US$100k+ (earlier only $20k) of seed capital, working with a portfolio across multiple domains, and ending after three months with a demo day. The best I can gather after talking to different programs around the world is that they go with this model because that’s what everyone else does. That really mean that what originally worked for Y Combinator (who originated the model in 2005) led to everyone else copying that model instead of innovating from there as a starting point.

Running Startups Unplugged (as a bootcamp) and then AcceleratorHK as the first accelerator in Hong Kong meant that the door was open to experiment with different formats. And by taking the time to understand what was needed in Hong Kong and what would help the entrepreneurs who traveled in from other countries, we built something that fit those needs. We were more hands-on than we would have been in the US, worked out of a co-working space to allow visitors to come and go easily and focused more on building sustainable companies with limited local investors than we might have elsewhere.

Of the many accelerators out there, lack of funding, lack of direction, lack of engaged mentors, or lack of people to run it can all shorten the lifespan of any program. They will close, limiting the strength of an alumni network, but their brief existence is still better than having no program at all.

Of the thousands of startups that go through accelerators, most will not survive past a couple of years after the program. Even for the most elite accelerators, success rates are far from certain, once you dig into the numbers instead of taking them at face value.

There are many paths to entrepreneurship and discovery. You don’t need to run a startup to qualify as an entrepreneur or to learn. While a lot is written about what I’ll call “traditional startups,” there is a much larger group of people already doing innovative work in areas that get much less attention. These are people for whom the de facto models don’t work. Research-driven startups that have longer times to market, startups where the team doesn’t fit the mold (being older, lacking educational credentials, working in unsexy markets) — they mostly ignored by accelerator programs.

Accelerators are usually just a first step

Often when I meet investors based in tech hubs, they say that there are too many accelerators — too many to choose from a limited pool of strong, investable companies. Often when I meet startups, at least early-stage ones, they are interested to apply to an accelerator — while also facing acceptance rates of low single digits. When I talk to people outside of the tech hubs, they sometimes bring up developing an accelerator as a way for their city to distinguish itself. When I talk to the people who run these programs, they often believe that they’re doing work that is good for their communities.

So there are four motivations here

1. Investors need to show returns and therefore are interested in accelerators insofar as they provide potential investments, which they do less often. While volume goes up, attractiveness of investment may not keep pace. Demo days crowded with other investors are not the best place to find early-stage startups to invest in. However, these investors will often support programs by mentoring and often support the work that is being done.
2. There is plenty of demand from entrepreneurs for accelerator programs. Judging from application numbers and acceptance rates (often in the low single digits), there is excessive demand on the part of entrepreneurs. Investors tell me that most of the startups applying are not good investments and therefore these low numbers are appropriate. But a better question is, what is the purpose of the accelerator? Who is it there to serve?
3. There is more than one reality in the world of startups. The Valley may be the major contributor but life outside of the tech hubs is very different. Starting an accelerator can distinguish an off-the-beaten-path location, attract talent and potentially keep talent local. I saw that while running AcceleratorHK in Hong Kong. I’ve heard it from people in other cities too.
4. People who run accelerators want to help the startups. They care less about the money and more about the impact. This might be long-term impact on their startup communities or it might be short-term impact by helping individual entrepreneurs build their businesses. There are much easier ways to make a living than running an accelerator, just like there are easier ways to make a living than running a startup.

Most programs lack industry focus, mentors with specific domain knowledge, and follow the typical three-months to demo day format.

There are cultural and stylistic differences between every accelerator but apart from that, few construct a program based on a theme, an industry or technology.

I’ve seen the same celebrity mentors listed on many programs. How involved are they? It’s not that they don’t want to help, but how much time can they really devote when they’re listed on multiple programs and have other work to do?

Is the typical three-month program length appropriate for what you want to do? If other programs didn’t do it first, would you have chosen that length of time yourself? Why? Three months goes by in the blink of an eye. If you allow your startups to change course in month three — which you absolutely should do — then what happens when the program ends a few weeks later?

What support do you offer your “graduates” after the program ends?

Do you really need to do a demo day? Or is it a distraction? What’s the goal for your demo day, beyond keeping people focused on an end point?

Do your startups have a fighting chance to raise money after the program? Or should you focus them on surviving by building a sustainable business in a market without investors? That’s what most of the world looks like, by the way.

When you think there’s only one way, you miss opportunities.

Stop trying to build another Y Combinator and build something that works for you. Markets are different, people are different and there are many ways to run an accelerator. The traditional model should not be duplicated everywhere. But we act like it does apply everywhere. Actually, we act like the Hollywood version of the traditional model applies everywhere. That’s what happens when you read TechCrunch too often.

So I want to support a different model: the External Innovation Model.

This is how it works for institutions (since this is not limited to for-profit companies).

The institution sponsoring the accelerator issues a set of challenges to which they need solutions. They collect applications, but judge the applicants on their ability to build, not on their ability to market or present themselves. Over the year (there is no need for the entire cohort to start and stop at the same fixed dates), a program lead manages the entire process and institutional mentors guide the startups. There are in-person group meetings every six months to show progress and to make go/no go decisions on continued involvement or granting extra resources.

For a modest budget, this institutional sponsor could fuel a program with ten individual startups working on issues related to its success, travel for the periodic meetings and someone to run the program.

Purpose and Execution.

• Lower investment risk. Early on, ten small bets can be better than one large one. There are reasons why institutions can’t quickly do many small projects themselves, but having ten external teams devote their own time is not problematic for the sponsoring institution. Let the startup develop their solutions with guidance from the sponsor rather than develop solutions individually and then pursuing a sales relationship.
• Difference in opportunity cost. The ten startups gain even if their ventures are not successful. They learn, get access to some financing and make connections with the sponsoring institution. It would be prohibitively costly for the sponsor to support ten projects internally.
• No Demo Day. There are good and bad things to demo days, but this type of a program doesn’t benefit from a demo day in the ways that general startup accelerators do. First, there is not necessarily an outside investor that has either the interest or the ability to invest in these projects or companies. Second, there’s no need to include a large public audience for the purposes of outreach or brand building. Actually, some of these demos might bore people in the startup world to tears. And that is a good thing.
• Resources. Like accelerators, the company needs to invest a bit of capital into each startup. Unlike most accelerators, I don’t believe the corporate sponsor needs to provide workspace for or relocate the admitted companies to a central location. The exceptions to that may be hardware or biotech based programs where equipment access is a key differentiator.

Phase I: Define the problems to solve for this program. Admit startups into the program, have one central person work with them while bringing in mentors from the institutional sponsor. Define an end-point at which the different startups present back to the sponsor.
Phase II: The sponsor determines which startups to continue to invest in. The program continues, possibly in-house.
End game: Acquire technology, acquire team, or end involvement with the startups. The end game is different in this model because the entrepreneurs themselves probably want different things than those that apply to the traditional accelerators. Not everyone wants to build their startup the traditional way.

If you liked this post and want to discuss it or tell me why it would or wouldn’t work for you, you can reach me here.