Looking back — 1.5 years since “Copying Y Combinator”

It’s been nearly one and a half years since I originally wrote my paper on seed accelerators: “Copying Y Combinator: Why and How”, which focused on how other people or organizations could create their own programmes. I wanted to reflect on what has changed, and what hasn’t changed since, and what that means for the future.

To be clear, my viewpoint is looking at what it takes to make a seed accelerator successful. Think of the VC business model: not every VC fund is successful. (In fact, the median net return to VC fund investors has not been positive since 1998.) To be a continued successful seed accelerator program, you need to have a financial model that works, provide value to the companies that you invest in, and invest in the best possible companies. That means the best possible companies need to prefer your program to any other. There are tiers involved, and I’m interested in what makes top-tier results possible.

What has changed:

Y Combinator provides the most money to their funded companies

The deal with The Start Fund was huge news. There is now an investor that is willing to put $150k into every Y Combinator startup, sight un-seen. Many commenters have suggested that this will open up YC to older applicants, who will have enough income security to risk quitting a more senior job to do the program. Though most YC, TechStars and Seedcamp companies raise external funding anyway, having this publicly committed before you even start makes it a nice security blanket.

Perhaps more importantly on a tactical level is that it gives YC startups more breathing room in the fundraising race. Tommy at CarWoo wrote a great post about this. As he wrote:

[Fundraising] was a huge issue on the minds of all the YC companies and I know for a fact that it was a distraction. I know it was for us.

Finally, this is a deal for Y Combinator startups alone, and puts them at a significant advantage compared to other seed accelerators in attracting the best companies to the program. (The highest-ranked factor startups think about when it comes to seed accelerators, per my original post.) The existance of The Start Fund signals that YC acceptance is a high bar, and it’s worth having an option on every company that comes out of it.

Y Combinator has the largest alumni network

According to my spreadsheet tracker, Y Combinator has funded just about 250 startups. Combined with an ethos of helping each other out, this is a huge advantage to potential applicants. I’ve spoken to a number of YC startups, and all of them mention this as a significant benefit. (This is the second-highest-ranked factor startups think about.)

The secondary effect of all this is that Y Combinator has seen 4x-20x the number of startups that other accelerators have seen. That means 4x-20x the applications, 4x-20x the founder problems, 4x-20x the customer acquisition problems, etc. They have more experience with pretty much everything. That extra experience is valuable for entrepreneurs.

TechStars has developed into a world-wide network

TechStars has recently announced a world-wide network of 17 seed accelerators, the TechStars Network. What’s interesting with this is that it spreads the business model of seed accelerators more widely, and starts to standardize on best practices. When I spoke with David Cohen nearly two years ago, helping other entrepreneurs and accelerators get started was clearly something he felt strongly about. (Which is different to the go-it-alone approach of YC.) While I believe my original thesis is correct, making each and every one of the accelerators that are popping up better is a great thing for entrepreneurs and startups.

What hasn’t changed:

Y Combinator is still the only seed accelerator in the Silicon Valley area.

There are new seed accelerators opening in what feels like every city, state, and university campus… except Silicon Valley. (More about this below.) I find this really strange, to be honest. The Bay Area is pretty much the richest source of technology startup resources, and most of the programs dedicated to the most fledgling companies don’t exist here? I can think of two reasons for it.

One, the people interested in starting seed accelerators want to do them in their own hometowns, no matter how suitable those cities are for these programs. This would explain why so many are started in other cities, but wouldn’t necessarily explain why no additional accelerators have been founded in Silicon Valley.

Two, Y Combinator is seen as an 800-pound gorilla in the seed accelerator world, and no one wants to get in a pissing match with them. This seems plausible, but I think there are so many resources in SV that any new program wouldn’t intersect with YC. Perhaps this is more reflective of an ego issue; that no one wants to start a separate program and then be compared to YC?

Y Combinator, particularly through Hacker News, is more directly engaged with startup culture

I’ve grown to think that Hacker News is a key differentiator between the different accelerators. It provides a strong conduit between the YC partners, the YC alumni, YC applicants, and general entrepreneurial people. These communities always have existed before (ie, Slashdot), but Hacker News has centralized the audience around internet startups, and more importantly around the Y Combinator experience, philosophy, and brand.

Hacker News helps feed the virtuous circle that makes Y Combinator a top-tier seed accelerator.

Most seed accelerators are just local copies of Y Combinator

I’m disappointed when I see a program that’s simply a clone of Y Combinator in a different city. Differentiate yourself! Though there are some notable exceptions:

TechStars has definitely taken a different tack with their program, developing into a network of international programs. Their core programs in Boulder, Boston and Seattle together have a higher level of experience and engagement, with more startups funded and close coordination between the different locations. TechStars has a mentorship process where each of their startups is matched with 1–2 mentors, and the mentors don’t work with any other TechStars company in that batch. (They also provide centralized office space for the startups.) All in all, it’s one of the most developed programs outside YC.

Fundamentally, results are what matter. TechStars actually publishes their results online, and they’re solid.

Seedcamp has a radically different approach, but probably befitting their non-US location. (Which means they can easily attract non-US startups that wouldn’t easily be able to live/work in the US.) They host numerous mini-Seedcamp events across Europe, and then cap it off with their final decision around the handful of startups they put more resources into. So while they’ve seen a lot of European startups, they’ve only invested in a fairly limited number, about 40.

The Brandery looked very interested when it got started. They’re located in Cincinnati, Ohio, which seems like it’s off the beaten track for startups… until you remember that it’s the world headquarters for Proctor & Gamble, a company that is incredibly focused on consumer brands. There’s a huge resource of talent for companies that need strong consumer brands.

I was hoping that this would open The Brandery up to startups that weren’t just consumer internet startups, but it looks like the list of their 2010 companies were just that. I’ve now become more dubious; at the early stage of these startups they need a product more than they need a brand, so the accelerator won’t be able to offer as much value. Perhaps there’s a place for a slightly later-stage startup: one that has a solid product but needs a brand finishing school to take them to the next level?


Andrew Parker (formerly of Union Square Ventures, now at Spark Capital) made an interesting comment on my original post:

If you’re going to copy YCombinator, then you should really give credit where credit is due: thank YCombinator. I went to the demo days for LaunchBox, TechStars, SeedCamp, fbFund REV and YCombinator in the past year. The only YC clone that even acknowledged that they were a YC clone and, furthermore, thanked YCombinator for their pioneering efforts was Dave McClure at fbFund REV. All the other programs never even mentioned YCombinator at the demo day.

Now I’m not convinced that every other accelerator should genuflect upon Y Combinator at each of their Demo Days, but what Andrew pointed out was interesting.

Y Combinator appears to be the least structured of all accelerators

TechStars was the second major seed accelerator out of the gate after Y Combinator, and they’ve started the trend of what appears to be much more structure in the TechStars program, and the programs TechStars has influenced. In addition to periodic meetings for everyone, they have all the startups work in the same physical office, and have a structured mentorship system. (As I understand it, each startup in a batch is matched with a small number of mentors, and those mentors work only with that specific startup.) This is quite a bit different from the YC model, where the major structure is a weekly dinner and then open office hours with the YC staff which you can take or leave.

What this means for seed accelerators:

You need to be unique, where unique is not just a seed accelerator in a different city

I am still absolutely convinced that if you’re a Y Combinator clone, just located in a different city, you will never be a top-tier program. Why? Because if you’re just doing exactly what YC does, but you provide less money and less expertise, you’ll never have the top startups wanting to work with you.

What I’m waiting to see is a program that does something else entirely. For example, what about an accelerator that works with companies building actual, physical products? Companies like Wakemate (from YC) have struggled as they work out production issues, find and develop relationships with factories, etc. If there was an accelerator that had a group of mentors that could help guide startups through this journey, with contacts a low- and higher-rate production facilities, and in a financial structure that made sense, I think it could be tremendously successful. (Again, defining success as being the top-choice of any startup working in that field.)

This is simply one example; you could do a dramatically different approach any number of ways. (After reading Roger Ehrenberg’s blog, couldn’t NYC start one around a theme of big data, data visualization and finance?) But so far, everyone just wants to have an accelerator for internet software startups, generally consumer focused, with the same model as YC. There’s more potential out there, people!

There’s room for another seed accelerator in Silicon Valley

Who’s going to start it?

(Dave McClure seems to have made a stab at this with the 500 Startups Accelerator; the main difference from others being that there’s no open application, startups have to be referred in.)

You need to be a program that everyone you’re focused on badly wants to attend

As one famous Google executive says, “Repetition doesn’t spoil the prayer,” and thus I want to keep repeating myself here. To be a truly successful seed accelerator, you need to be highly desired by the best companies you want to help. There needs to be a strong match between what you offer, and what the startup wants and needs to be successful. There will always be companies looking for investors, and if your pool isn’t that big some of them will certainly look good, if only by comparison. But your accelerator needs to be the preferred program for the best startups. Those best startups have the best chance of being successful and generating the results and returns that enable the program to continue sustainably.


I still feel like we’re in the very early days of seeing the successes and failures of seed accelerators. The startup world, and specifically the funding world, appears to have a number of discontinuities. (Particularly once you start getting away from consumer internet startups.) Seed accelerators have a great opportunity to start filling in the existing gaps, and helping companies go from idea to polished execution much more cleanly.

Going back to a link I shared near the top, Bryce Roberts has some great comments about VC funds that I believe also apply to seed accelerators:

a handful of them have been delivering outsized returns for decades now. They don’t call Sequoia, Accel, Benchmark, KP, Matrix, Greylock “top tier” for nothing. They’ve figured out a few things related to building enduring companies and consistently delivering returns for their investors.

If you’re going to start a new fund, be different. Proprietary dealflow, investment stage, operating experience or deep network of industry contacts are meaningless buzzwords that aren’t going to set you apart from the pack. As SuperLP says “to do something outstanding takes audacity. And indeed, private equity should be all about audacity”. Being the 10th seed fund, or 5th “opportunity” fund isn’t going to set you apart from the pack. Be different.

To create a top-tier seed accelerator, you need to be a top-tier choice for startups the world over in your niche. Just like a VC firm, seed accelerators need to have the best possible startups (deal flow), in order to fund the best teams and ideas. If your seed accelerator can achieve this, you will have a sustainable program.

Appendix — Data

If you haven’t seen it already, I’ve maintained a list of companies that have been funded through these seed accelerators. Click here for the Google Docs link, or check out the embedded doc below:

%d bloggers like this: