The Kauffman Foundation has some data on angel investor investment and returns at their Angel Investor Performance Project. The data has some serious limitations, mainly that it is relatively concentrated in time (44% of the investments with data are ’99 or ’00), that there’s not enough investments to draw fine-grained conclusions, and that valuations at each investment are not available so follow-ons are hard to interpret. That said, data is better than no data. A summary below*.
| Multiple | |||||
| 0x | 0x-1x | >1x | All | ||
| No follow-on | |||||
| Total companies (% of all) | 97 (66%) | 50 (61%) | 158 (81%) | 305 (72%) | |
| Average exit multiple | — | 0.47 | 28.27 | 14.72 | |
| Average years invested | 2.9 | 2.8 | 3.6 | 3.2 | |
| Average invested | $88,420 | $109,678 | $90,125 | $92,788 | |
| Follow-on | |||||
| Total companies (% of all) | 50 (34%) | 32 (39%) | 37 (19%) | 119 (28%) | |
| Average exit multiple | — | 0.33 | 7.00 | 2.27 | |
| Average years invested | 2.8 | 4.5 | 4.9 | 3.9 | |
| Average invested | |||||
| Initial | $346,460 | $86,836 | $129,730 | $209,259 | |
| First follow-on (% of initial/% of companies) | $194,345 (56%/100%) | $41,946 (48%/100%) | $92,524 (71%/100%) | $121,705 (58%/100%) | |
| Second follow-on (% of initial/% of companies) | $66.080 (19%/20%) | $17,425 (20%/13%) | $13,097 (10%/8%) | $36,523 (17%/14%) | |
| Later follow-ons (% of initial/% of companies) | $19,600 (6%/10%) | $5,625 (6%/13%) | — (0%/0%) | $9,748 (5%/8%) | |
| Total invested (% of initial) | $626,485 (181%) | $151,831 (175%) | $235,350 (181%) | $377,234 (180%) | |
| All | |||||
| Total companies | 147 | 82 | 195 | 424 | |
| Average exit multiple | — | 0.42 | 24.23 | 11.23 | |
| Average years invested | 2.9 | 3.5 | 3.9 | 3.4 | |
| Average invested | |||||
| Initial | $176,189 | $100,764 | $97,640 | $125,477 | |
| Follow-ons (% of initial/% of companies) | $95,247 (54%) | $25,364 (25%) | $20,041 (21%) | $47,144 (38%) | |
| Total invested (% of initial) | $271,435 | $126,128 | $117,681 | $172,621 | |
Some thoughts. The data is noisy and there are outliers that skew it. This shouldn’t surprise anyone. I would have expected that the deals with follow-ons would have returned more absolute dollars than those without, albeit at a lower multiple. The lower multiple is there, but the total average dollars returned is about a third. This means one of two things: 1) angel investors primarily invest in follow-ons when the company sucks, or 2) there is not enough data to smooth out the statistical anomalies. I think the second is probably truer. But there is also some truth in the idea that angel investors may not get a chance to invest in later rounds when the company is going gangbusters. The deals with follow-ons also exit later, even the successful ones, so some companies may not have angel investor follow-on because they exit before another round is needed.
Failure rates are pretty high. About a third of the companies fail outright. Another 19% don’t return capital. The remaining 46% pay out. Here’s a graph of exit multiple vs. years invested.
I cut off one or two multiples above this range and a couple of exits that were more than 14 years out so you could see the structure. I would describe this as consisting of two things overlaid (although I don’t think there’s enough data to prove it): 1) a group of investments that follow a constrained exit pattern–this is the bulk of the exits, peaking at three years and then declining until it disappears at eight years–and 2) investments that don’t seem to follow a particular pattern. This jibes with my gut that says some investments are ‘earners’, they are less risky but have less upside. These companies build then exit. Other investments are the ‘swing for the fences’ types. These can exit anytime and for any amount.
Here’s a chart of how long the angel investors held their investment before exiting.
And here’s a chart of exit multiples.
This is a log-log chart, showing that even with this noisy data, exit multiples follow a power law.
—–
* I cleaned the data. I removed any deals that: did not have exits, did not have a number greater than zero in the ‘totalinvested’ column, and any where the ‘stage’ field was not either blank, seed, or startup.