New Seed-DB research published

Seed-DB has been a member of the Accelerator Assembly since the organization began last year. The Accelerator Assembly is a “Startup Europe” initiative funded by the European Commission, but importantly is an industry-led network. It was founded by seed accelerators in Europe (Seedcamp, TechStars/Springboard, Bethnal Green Ventures), as well as Startup Weekend, Seed-DB, and several others, with strong support from Nesta.

One of the tasks that this group of organizations agreed to was to do two pieces of research. The first was on the overview of the seed accelerator ecosystem, particularly as it applied to Europe. The second was on the results and impressions from the startups that graduated from accelerators. Seed-DB (and specifically me) was tasked with combining quantitative analysis, largely based on Seed-DB data, with qualitative research based on surveys of accelerators and startups. THANK YOU to all of the accelerators and startups that answered the surveys… there is clearly survey fatigue amongst accelerators and startups world-wide, so I appreciate your time.

Both reports can now be found and downloaded from the Accelerator Assembly research page.

These reports are somewhat limited in scope, because the European Commission has very specific items that they required us to study. Despite this, I think there are some insights that may be useful for accelerators and startups alike:

Seed Accelerator Ecosystem paper

  • Size of accelerator ecosystem: The European accelerator ecosystem makes up a smaller proportion of the number of total companies funded (20.3%) than the number of programs world-wide (27%). This is because all of the programs that accept significantly larger class sizes (like Y Combinator, 500startups) are located in the Americas.
  • Size of job impact: Based on some reasonable assumptions, the number of jobs shown on Seed-DB is likely only 50% of the total number of jobs created by seed accelerator graduates.

Startups paper

  • The top two reasons that startups cite as important benefits are: 1) Mentorship / Coaching / Feedback, 2) Network / Alumni / Prestige. But one of the top drawbacks of accelerators cited by startups was also mentoring! It’s clear that mentoring is a highly valued aspect of a seed accelerator program, but when done poorly can be damaging to startups.
  • Another top drawback is the focus on Demo Day. As seed accelerators proliferate, the number of interested/qualified investors at each event will quickly diminish except for the top programs. Spending significant time on Demo Day when there’s no clear benefit to the startup detracts from the startups’ experience.
  • When asked how much equity a startup would be willing to give their accelerator even without any funding, nearly 80% gave non-zero answers, with 55% stating between 1–6% equity.

Please do have a look at both papers when you have some time. And as always, get in touch with me directly if you have any questions or comments.

Seed-DB in the Economist!, new feature launches, 2014 plans, and more

Happy New Year! In light of the Economist article this week that featured Seed-DB data, I’ve got a whole bunch of new features to share that just launched, some benchmarks on Seed-DB data, and plans for 2014.

Economist article

This week’s issue of the Economist features a special report on “Tech Startups”, and I’m happy to say that it used Seed-DB data in a chart to show a sample of result data from accelerators. It also cited Seed-DB data (and even mentioned my name) on the financial summary of the accelerator ecosystem in the section titled “Getting up to speed.”

2013 Achievements

Over 3100 companies and over 180 seed accelerators are now tracked on Seed-DB. They’ve raised funding of over $3.2billion and have created well over 11,000 jobs. About 25% of accelerators have administrative access to add companies and modify their pages themselves. (I’m hoping to increase this significantly in 2014.)

New Features on Seed-DB:

  • Top Investors League Table: Identify the top investors in companies that come from seed accelerators by the number of companies they’ve funded, the number of individual funding rounds they’ve participated in, or the number of different accelerators from which they’ve funded graduates. Top 12 of each list are shown on the investor graph page.
  • Administered entries: Accelerators that manage their own page on Seed-DB now show up with a yellow dot on the list of all accelerators. The name of their page administrator(s) are shown on the accelerator’s page.
  • Tables of funding, exits: I’ve long had a feature to list startups that have raised a particular amount of money (or more), and a feature to list all companies that have exited. These are now more clearly presented in the new “Tables” menu. If you want to see companies that have raised more than a particular amount of money, just start from the list of companies that have raised >$1million, and replace the number in the URL with the value you want.
  • Clarified how individual funding rounds are displayed on each company’s page
  • Seed-DB should be running faster! I’ve improved data caching on the site, and made some performance tweaks
  • Improved the menu structure on the site to make it easier to understand/navigate
  • Investor data is improved, and should stay fresher over time.
  • You can help pay for Seed-DB running costs by donating bitcoin. (None of this will go to Jed; it will go solely to paying server bills and other incidental costs of running the site.)
  • The development roadmap has been updated
  • Progress charts. I’m now collecting daily summaries of the data on Seed-DB. Once I’ve built up a little more history I’ll be displaying how Seed-DB is constantly increasing the number of companies, the number of accelerators, and the amount of funding tracked

Benchmarking data

A lot of people ask me question along the lines of “how accurate is this data”? The answer is two-fold, since there are two primary sources of data.

The first data source is what companies have graduated from what accelerators. This data is very complete for the larger accelerators, less so for newer accelerators. I continue to work to make this as complete as possible, but some accelerators go out of their way to make it difficult to find out what companies they’ve funded. (And some accelerators make it difficult for me to determine if they qualify as an accelerator.) In general, this information may be incomplete in places, but it is rarely incorrect.

The second primary data source is on funding. This comes exclusively from Crunchbase. Some companies have not listed their funding on Crunchbase, and this is the primary reason why Seed-DB funding data is lower than what some programs advertise. Reasons for not listing data range from strategically staying quiet about resources, to not caring much about what’s shown on Crunchbase. Per Paul Graham’s tweet this week, we can calculate that Seed-DB captures ($1.748bn / $2.09bn) = 84% of the total YC company funding. I’ve been informally told that TechStars has a similar factor. Again, the data is as complete and accurate as the individual companies make it.

2014 goals
I have a few key goals for Seed-DB in 2014, and would welcome your feedback and suggestions on more ways I can develop the site. Here are my plans/goals:

  • Increase ability and frequency of accelerators to manage their own pages on Seed-DB
  • Completely refresh the look-and-feel of the site
  • Integrate more statistics for analysis, such as the markets for each startup from their Angellist entries
  • Make upkeep of Seed-DB less manual and more sustainable

Final note

On a personal note, I’m looking forward to focusing a bit more time on Seed-DB in 2014. For those of you that don’t know, it’s just a side project of mine… my day job is at Google in London. Last year was tough for me to spend much as much time on Seed-DB as I wanted; my daughter started walking and going to day care, my family moved to a different home, and my day job was absolutely crazy. (I’m currently hiring three people for my team!)

Someday I still plan to write the story about how I taught myself Python/App Engine, built Seed-DB, became a father, and launched a bunch of interesting new features all in a ~nine month period. That’s a good one…

Announcing the Seed-DB Investor Graph, and responding to the ‘VC Zombie’ s**tstorm

  • Seed-DB Investor Graph: EXCLUSIVE first look
  • “Zombie VC” s**tstorm & Crunchbase data
  • TechStars expands by merging with existing programs

Seed-DB Investor Graph — EXCLUSIVE first look

I’m pleased to announce the launch of the Seed-DB Investor Graph! The Investor Graph shows the web of VC and angel funding, by detailing who has funded which accelerators/companies as well as which round (seed, a, b, etc.) and when the funding occurred.

Diving into the data of individual investors, you can see that Google Ventures has funded companies from a number of different accelerators. In contrast, Andreessen Horowitz generally only funds companies from Y Combinator, though with some exceptions.

From a different viewpoint, entrepreneurs might be interested in which investors actually fund companies from different accelerators. (Not just who shows up to the Demo Days, for instance.) In the Investor Graph you can see that Angelpad has a fairly diverse group of investors that have funded their companies. Though I’m not going to name names, this tends to be a very different story for newer/smaller accelerators.

Finally, the individual company pages have now been upgraded to show information on funding rounds. For example, Songkick shows their progression from a Y Combinator investment in 2007 through to further investment by Index Ventures in 2008 and a large investment by Sequoia Capital in 2012. Alexa traffic rank charts and Compete unique visitor charts have been included.

The Investor Graph is still in its earliest stages. There are still some intermittent bugs I’m trying to quash, and there is a lot of functionality and visualization that I could potentially build. But in the spirit of a minimum viable product, please let me know what would be useful to you. (Or if you experience any bugs! Though typically re-loading the page resolves the error.)

“Zombie VC” s**tstorm & Crunchbase data

Danielle Morrill of Referly has recently been on a blogging tear, and in the meantime been kicking up a bit of a s**tstorm in some VC/Angel circles. It all started with her excellent post about zombie startups. (A future project for Seed-DB is to better identify potential dead/zombie startups._

But last week Danielle posted about zombie VC’s which seem to have touched a nerve for two reasons. One — she programatically went through Crunchbase to identify active vs. inactive VC’s. This didn’t work because of Crunchbase’s data, which I’ll discuss further below. Two — she called out the zombie VC’s instead of focusing on the active VC’s. Focusing on the negative just opened her up to further sensationalism, though I’m guessing it certainly drove pageviews!

Through Seed-DB I’ve been deeply enmeshed in Crunchbase data for the past year, so I think I have some unique insights. The fundamental thing about Crunchbase is that:

Crunchbase data is largely accurate, but often incomplete.

Where people have taken the time to contribute information/data, the entries are pretty good. But so, so many companies (even companies with good funding) simply don’t have data in Crunchbase. The biggest sin I’ve seen in Crunchbase data is duplication, of funding rounds, of investors, even companies. For example, Danielle’s company has two entries in Crunchbase, and the active/maintained entry has duplicated funding information which makes it appear that has raised two $1million rounds instead of just one $1million round. For Danielle, when it came time to determine algorithmically which VC’s were “zombie” VCs, the incompleteness of Crunchbase meant that active VC’s were mistakenly identified as “zombies”. But in a brilliant marketing strategy, the sensationalism of the headline and initial results meant that a lot of these VCs were very… motivated… to get in touch with her to correct the results!

Like Danielle, I think Crunchbase is an absolute treasure for the startup community, but data from it needs to be understood in context of “largely accurate, but often incomplete”.

TechStars Expansion

TechStars has been growing since it established itself in Boulder in 2007. However, creating a new program from scratch in a new city certainly requires significant investment, both in cash, facilities and most importantly people. Over the last 3 months, TechStars has expanded in a big way by merging with existing successful seed accelerators; Excelerate Labs is now TechStars Chicago while Springboard is now TechStars London. Expanding this way has interesting benefits for both parties.

Excelerate Labs and Springboard were both running very well-regarded accelerators in Chicago and London/Cambridge, respectively. But with over 150 different accelerators around the world, it’s difficult to rise above the fray. Merging with TechStars linked Excelerate & Springboard to a large and growing brand that can help drive even higher quality dealflow to and through their programs. Additionally, TechStars has developed expertise and an organization that’s engineered to support startups regardless of location or program.

By merging with existing programs, TechStars has removed the risk and limited the investment necessary to expand. Since TechStars runs/fosters the Global Accelerator Network, popular amongst accelerator program administrators, they naturally have developed relationships that will allow them to do this more easily in the future. It is quickly making TechStars a well-known institution around the world, and I’m sure having positive benefits for TechStars dealflow.

I suspect what TechStars is doing now would benefit other accelerator brands as well and we could see a new trend of brand consolidation emerging in the next 1–3 years.

“The Launch Pad” review, thoughts on scaling accelerators, and important new Seed-DB features

Review — “The Launch Pad” by Randall Stross

What is it really like to be involved with Y Combinator? “The Launch Pad” is the newest and most complete piece to date. While other writers have penned articles about YC (notably Stephen Levy in Wired magazine), this is the first book. It follows the Summer 2011 class of Y Combinator from the applications/interviews, during the program and Demo Day, through to the successes and failures the startups have experienced to date. In addition to this class, the author has interviewed a variety of founders across numerous YC classes to provide a more detailed history.

If you don’t know much about Y Combinator, I would highly recommend this book. It’s a very thorough picture of what it’s like for the founders of the startups going through YC (including what it’s like to get accepted!) as well as insights into the YC partners’ experience. It details a microcosm of the startup experience; initial successes, technical breakdowns, pivoting business models, and debates about age and gender. I found it a nicely complete book.

For people who are familiar with Y Combinator, you will probably enjoy it but you will likely not learn much more than you already know. There were some details that I hadn’t heard before, like the exact investment splits of the early YC partnership. And there are a lot of interesting stories and anecdotes about a lot of different YC founders and YC partners. But there’s little that would be unfamiliar or groundbreaking if you’re a regular Hacker News reader.

Overall, if you’d like to get more detail and color into the Y Combinator experience, check out “The Launch Pad.”

Thoughts on scaling seed accelerators

One of the more interesting phenomenons I see in seed accelerators is how they scale, and in general I see three models. There’s the YC model, which features large class sizes (>25 startups per class) and/or continually adds more and more startups to each class. There’s the TechStars/franchise model, which keeps class sizes largely the same (5–15 startups per class), but scales by adding more classes per year and/or more franchises in different cities. Finally, there’s the quality (aka default) model, which doesn’t change class sizes or frequency, but “scales” by getting better startups through the doors. (The last model is effectively just keeping the status quo and not scaling; I’ve included it here to be complete.)

There are only a small number of accelerators that use the YC model: Y Combinator, 500startups, and Startup Chile are the top examples. They have developed programs which scale with more startups, and through portfolio theory should have a better chance of generating a solid financial return on their investments.

Many accelerators use the TechStars model; TechStars in particular has popularized it through their Global Accelerator Network (GAN) with which ~50 accelerators are affiliated. In this model class sizes are “no more than 10–20 companies at a time” involving “40–80 mentors”. By capping the size of each class, it does limit the scaling potential; instead of adding more startups per class these programs add classes or add similar franchises in new locations. Clearly the GAN has taken a stand that it’s more important to provide the in-depth, personal mentoring for all startups in their programs than it is to scale to larger numbers in a class.

Finally, a lot of accelerators are really in the default model; these programs either don’t have enough quality deal flow or cash to invest to expand the number of startups they fund. In this case, they need to scale the quality of the startups they do fund by continually marketing their program so they can raise the bar for admission.

Which is the best of these models? There’s no definitive answer. I’ve discussed this with a number of people and currently believe it depends on the startup. Susan Cohen, a founder of and a PhD researcher of seed accelerators at the University of North Carolina, was particularly helpful and shared with me her thoughts from her research to date. If you’re coming into a program with just an initial idea, the programs that provide in-depth mentoring can help founders iterate on their ideas and business models faster because of the intense mentoring. But for startups going into a program with a well-established seedling business, the YC model can provide the time/space to hack and iterate on a product because of the lighter-touch mentoring. I’d be interested in your thoughts; please send me an e-mail with yours!

Seed-DB three big new features

I’m very happy to officially announce three big new features in Seed-DB. Some of these have been live for a while, but together they make the site much more useful.

1- Accelerator ownership

I’m proud to announce that accelerator owners can now get access to edit the entries of their accelerator and the startups they fund on Seed-DB! Access will be granted manually to verify that only program partners/administrators are given access. But once granted, you’ll be able to edit your own information, add startups you’ve funded, and provide information on when your program is open for applications. To get access to your program, follow the instructions on the guide page for accelerators.

By doing this, I’m hopeful that the data on Seed-DB will be even more up-to-date, since the program administrators will be able to make any changes without me (Jed) being a bottleneck.

2- Application dates

Accelerators can now provide information on application dates to their programs! Simply provide the date applications open, the date they close, and a link to the application page. If your program currently is open for applications, it will be highlighted on the list of seed accelerators. And if you’re a startup wondering which programs are currently open, just check out a single page:

I hope this will help accelerators better market their programs to interested startups, and it will help startup founders discover and apply to as many programs as they fit.

3- Funding over $X

Have you ever been curious how many startups have raised more than $X in funding, where X is a value that you care about? Well now you can. To get the information, use the URL in the form: where X is the funding level you care about.

For example, as of today there are 237 companies in Seed-DB that have raised >=$1million in funding: 64 companies have raised >=$5million and just 28 have raised >=$10million.

Finally, to see data on all the companies that have achieved an exit (currently 102 startups), simply check out this page:

Please remember that the “H” / “M” / “L” letters next to the exit value column represent the Seed-DB confidence level in that particular value. “H” is reported/verified information, “M” is rumored value, and “L” is an educated guess.


Thank you for reading to the end of this post. I really value feedback; please send me a short note or comment below on what you liked, what you didn’t like, or what you’d like to see more of in this newsletter. And if you liked it, please forward it to a friend!

Three new features in Seed-DB

I’m pleased to say that Seed-DB now has three new features! Some of these come from direct feedback; others are features I’ve been planning to add for a while. I hope you find them useful.

Median + Average funding

Brad Feld got in touch with me to refer me to his post where he discusses the differences between average and median funding in venture capital. Brad’s post makes a lot of sense for Seed-DB, too. I had previously just listed average funding for the startups in an accelerator; I have now added a column for median funding of the startups in that accelerator.

There is a slight hitch in how this is implemented because some accelerators have very few startups with Crunchbase profiles. In these cases even the median figures would be inaccurate because only the startups with solid funding rounds were typically shown. So if a startup doesn’t have any Crunchbase funding listed, I calculate the median using a combination of Crunchbase profiles (where available) and the average funding per startup (for each company that doesn’t have funding shown).

The result of this is that median values are accurate where most startups in an accelerator have up-to-date Crunchbase pages. TechStars in particular excels at this. As usual, if you have questions or concerns about this e-mail me directly.

Application Dates now available

There are over 120 seed accelerators listed on Seed-DB; this is a lot of choice for entrepreneurs around the world! As of today, Seed-DB will highlight seed accelerators with open application windows and direct visitors to the accelerators’ application page. These windows can be entered well in advance. On the day applications open a bright yellow button will appear next to your accelerator.

Only a couple of accelerators currently have entries. If you’d like to add information on the application open/close dates for your accelerator, enter them in this form. (I’ll check this link periodically and add what is submitted.)

I hope this helps startups learn what programs are open for immediate application, and helps accelerators generate additional application/deal flow.

Significantly improved updating

While not directly visible by users, the automatic updating of startups and accelerators has been drastically improved. Most changes in Crunchbase pages should be reflected in Seed-DB within 24 hours.

Seed-DB: a month of changes

Seed-DB was launched on July 25th, and since then it’s been a fascinating ride. Here is the latest news on everything related to seed accelerators and Seed-DB.

Follow Seed-DB on Twitter

Seed-DB now has a Twitter account: @GoSeedDB. Follow it to get all of the latest updates and announcements on the site. Like this newsletter, it will be low-traffic with just important updates for startups, accelerators, and investors.

Y Combinator Demo Day — public/stealth analysis

Some accelerators (like TechStars) publicly announce all of their companies at their Demo Days; while startups in other program (like Y Combinator) only publicly announce when they’re ready. Y Combinator’s Demo Day is taking place on Tuesday, August 21st.

In this years’ Y Combinator Summer 2012 class, there are 84 startups. As of today (Monday, August 20th), 50 of these startups (60%) have already gone public. I would expect another 10–15 startups to go public on Demo Day, leaving 20–30% of the startups to go public at some point in the future.

This past month on Seed-DB

I’ve focused effort on Seed-DB in the last month on two main themes:

  • improving breadth and quality of data
  • bug fixes in the application

In less than 30 days, the number of companies in the database has increased 25% to over 1600 companies. VC funding tracked has also increased by about 25% to over $1.28billion. And the number of jobs also increased 25% to over 3700 jobs created. Thanks to lot of feedback from accelerators and startups, I’ve cleaned up the data sources to improve the quality of data on Seed-DB, too. (There’s a link on every page where you can get in touch with me if you have data to correct!)

One problem with articles like Jared Konczal’s in Forbes is that the data in Seed-DB is still very preliminary; there are many gaps. I’ve tried to make these clear; users can now see what startups are associated with Crunchbase pages (and thus I’ve able to get funding/employee data) and which startups don’t have this data.

Finally, by launching and getting valuable feedback from a number of people I’ve identified and fixed a number of bugs in the application. Thank you to everyone that’s reached out to me with corrections.

Important upcoming feature

The biggest upcoming feature is an application deadline tracker. I’ll be adding a column to my list of accelerators to highlight programs that are open for applications. I hope this will help startups find accelerators that they might not otherwise know about.

Seed accelerators in my list will be able to send me the date applications open and close, as well as a link to their application page. If you run a seed accelerator, please e-mail me and I’ll add you to my announcements list once this functionality is ready to go.

Advice for seed accelerators

There are a few recommendations I have for seed accelerators everywhere:

  1. Make it easy to find your startups; a portfolio page or something similar. (You’d be surprised how difficult it is to find startups from some programs.) If you can’t promote your startups on your own website, why are you doing it? Ideally categorize them by class.
  2. Get startups to create a Crunchbase entry, and at least ensure they add the funding they got from your accelerator.

Thanks for listening; and be sure to follow @GoSeedDB!

Introducing Seed-DB: a database of seed accelerators and their companies

I’m pleased to announce Seed-DB, a database of seed accelerators and the companies they have accelerated. I originally started assembling this data in the summer of 2009 as part of the research paper I wrote for my MBA thesis at the University of Cambridge: “Copying Y Combinator: a framework for developing Seed Accelerator programs.” Ever since, I’ve been keeping it up-to-date in a number of Google Spreadsheets.

Seed-DB puts all of the best information I have and puts it in a much more digestible format for easy viewing and comparison. I have integrated Crunchbase data to get up-to-date information on funding rounds, number of employees, and more. I believe that Seed-DB is the most useful and comprehensive resource on seed accelerators around the world.

Economic impact

One of the clear things that stands out to me is the economic impact of seed accelerators. Even though the data is far from complete for some programs, these are some key statistics as they stand today:

  • >100 programs world-wide
  • >1300 companies have been accelerated
  • ~65 companies have already exited for an estimated $930million
  • >3000 jobs created
  • >$1billion in angel and venture capital funding raised

In particular, as the data becomes more complete I estimate that the number of jobs created could easily double. And as the companies mature and have exit events for investors, the exit values will be much, much larger. (Paul Graham has publicly stated that just the top 21 Y Combinator companies have a combined value approaching $5billion.)

Key disclaimers

Incomplete data

As hard as I’ve tried, the information on Seed-DB is NOT 100% complete. There are three reasons for this:

  • Missing startups. Some accelerators have made it very difficult to find lists of startups that have gone through their programs, so there are many startups missing in less visible programs. Please contact me if you can provide me with any details.
  • Missing data in Crunchbase. Many companies have not entered any information in Crunchbase. That’s where I get details on funding and number of employees, so if you want this information to be accurate please update it there. (If you run an accelerator, it would be ideal if you could encourage your startups to do this.)
  • Missing accelerators. I believe I’ve tracked down all of the seed accelerators world-wide, but please let me know if I’ve missed some. Note that I’m using a pretty strict definition of a seed accelerator. I hope to eventually track other types of programs, too.


To use a buzzword, this is a minimum viable product. Reid Hoffman has been quoted as saying “if you’re not embarrassed by the first version of your product, you’ve launched too late.” Well, I’m at that point. There’s a lot that I still want to do, some of user-visible projects and tasks are on the roadmap, but what’s launching today is the core of my vision.


Please note that while exits are often publicized, the price paid for most companies is NOT public. So for the vast majority of companies I’ve simply had to guess. I’ve put a confidence level by each guess, with the highest-level confidence reserved for the cases where the price was public or widely reported. These values do not come from any of the companies involved.

Final word

If you’ve made it this far, I’d ask you for three things.

  1. Sign up to my seed accelerators newsletter. (It’s very low traffic, and I don’t share or sell e-mail addresses… ever.)
  2. Send me feedback. I want to make this useful for startups interested in accelerators, people founding/running accelerators, and investors interested in companies from accelerators.
  3. If you’re interested in collaborating with me on the technical side of things, please get in touch. I’m new to web applications, and could use an experienced partner.

Thanks for your attention!