Photo Credit

Accelerators vs Angels

An entrepreneur friend asked me if he should apply to an Accelerator or go raise Angel money. Generally speaking this isn’t an either/or sort of question, and what is best for each company is a function of its particular needs and situation. But, we can bring some useful historical data analysis to see how these paths correlate with good near-term outcomes.

A good early indicator of “success” is whether or not a subsequent Series A financing happens for the company. This older post of mine — Do Angel Investors in California Outperform the Rest of the World? — examines the rationale and motivation for using this success metric. While a subsequent Series A is not any sort of guarantee of success, it is essentially a necessary condition for major success, and only happens for a modest fraction of all companies that graduate from Accelerators or receive Angel money.

Fortunately, all the data we need to analyze the correlation between graduating from an Accelerator or receiving Angel money and raising a subsequent Series A financing is available from the fantastic CrunchBase Dataset. The CrunchBase Dataset includes 96 Accelerators and Incubators and 3,462 individual Angels that made at least one investment in the years 2010 through 2013. The 2013 end was chosen to allow enough time for subsequent Series A investments to have had a chance to happen. Those 96 Accelerators and Incubators graduated 2,012 startups across those four years, with the largest half dozen programs producing about half of the graduates. Over the same period of time, the 3,462 Angels invested in 1,923 companies. There is likely some sample bias among the Angel investments, with the companies that ultimately fail never making their way into the Crunchbase Dataset. The year 2010 start of the consideration universe was chosen in part because that was about the time when CrunchBase was enough of a known information source that startup companies had an incentive to have as much of their funding information published as possible, which reduces the sample selection bias associated with Angels adding information after the fact, and potentially mostly for their winners.

The table below summarizes the results. We see that about 13% of the Accelerator and Incubator grads went on to raise a Series A financing, and about 23% of the companies that received Angel money went on to raise a Series A financing. The Standard Error on these success ratios are small enough that we can assume that potential sample bias issues outweigh any random noise.

A relatively modest fraction of Accelerator/Incubator Grads ever go on to raise a Series A financing. Success rates are better for those that raise Angel money, but the odds are still long. Consideration Universe Subset defined below.

It turns out that there were 484 companies in this combined population that both graduated from an Accelerator or Incubator and received Angel money. These 484 companies were more likely to receive a subsequent Series A financing than the 1,412 companies that received Angel money without graduating from an Accelerator or Incubator. The table below compares these two populations. The primary root cause of this difference appears to be the subset of companies in the second group who raise a very small amount of Angel money and quickly fade away.

Those companies that both graduate from an Accelerator/Incubator and raise Angel money have a bit higher success rate. This appears to be driven by the startups who raise small amounts of Angel money and never make much progress. Consideration Universe Subset defined below.

Looking more closely at the 2,012 Accelerator or Incubator graduates, we see three separate paths:

1. About 9% of the graduates go straight to raising a Series A financing.
2. About 24% raise Angel money. Usually, this happens at the time of graduation from the Accelerator or Incubator, or soon after. Of the 2,012 total graduates, about 7% reach a Series A financing via this path.
3. About 84% of the graduates have not raised a Series A financing. These companies have failed, been acquired (generally, for small dollar amounts), or are continuing to operate on a shoestring budget.

Conclusion

Every startup is different. For some, the Accelerator/Incubator path brings something they really need. For others, it is the only alternative other than giving up. But, it is only a step in a process, and that process is one is a long and difficult one. As I hope the data above illustrates, going through an Accelerator/Incubator program is not some sort of golden ticket to the promised land. Most startups fail, and the earlier the stage, the truer this is.

Methodology

I’m including this section for the small subset of folks who love to drill into the details. If you’re not interested in those sort of details, you won’t miss anything by skipping this…

The 3,451 companies in the consideration universe are the subset of companies in the CrunchBase Dataset that meet the following criteria:

• Primary location is USA. This represents the overwhelming fraction of the companies, and for this subset of the companies, the rest of the associated data has better coverage.
• Total financing size less than $2M. Financings larger than $2M don’t seem to me to be in the spirit of what most people would consider to be a Seed financing — at least, not for the 2010 through 2013 years.
• Round type = “angel”, “seed”, or “convertible note”. This removed a fair number of small “Series A” financings that would have otherwise met the initial selection criteria. The financing provided by the Accelerators and Incubators is generally categorized as “seed”.
• Investor was an individual Angel or Accelerator/Incubator. Explicitly removed were financings from Micro VC’s, traditional venture firms (many had seed programs in those years), government entities, and non-profits.
• This initial financing happened in the years 2010 through 2013.
• Have not already raised a Series A financing. A fair number of Angel rounds are bridges, extensions, follow-on financings, or small pieces added on to another round of financing. These may be great investments, but they are not in the spirit of the analysis I am trying to do here.

For a subsequent Series A financing to have been deemed to have happened, it had to meet the following criteria:

• Round Type = “venture”
• Round Code = “A”
• Total financing size greater than $2M
• Date has to be after (and not same date as) the initial date above, and was in the CrunchBase data set as of the September 30, 2015 quarterly update date. There are certainly a few companies that entered the consideration universe above that will someday reach a Series A financing, but haven’t done so yet. I would be surprised if this was even one percent of the universe, though.

This Wikipedia article on Standard Error, and this one on Bernoulli Distributions are a good refresh on the underlying statistics.