Is a Respectable ROI an Achievable Metric for Angels?

December 14 2009 No Commented

Is a Respectable ROI an Achievable Metric for Angels?

I’ve been funding startups as an angel investor since 1980 without having a good answer to this question.  Early stage VCs seem to earn over 20% IRR (annual internal rate of return), so my thought has been that angels should earn even more.  After all, we generally invest earlier with more risk than do seed VCs.  My own portfolio has done well, but one cannot base any conclusions on one lucky chap.

But now we have some answers.  Professor Rob Wiltbank of Willamette University has now concluded two studies on angel returns, one in the US (November 2007) and one in the UK (May 2008), all based on surveys of angels in groups.  Without reporting the minutia of each study, let me just say that each study included a large number of angels, angel investments and exits.  Angels in groups in the US enjoyed a 27% IRR while those in the UK saw IRRs of 22%.  Both are well within expected outcomes for these high risk opportunities.  Finally, we have some data for angels!

Wiltbank drew the following additional complementary conclusions in both studies.  The numbers reported below are from the US study:

  • Angel returns are quite skewed.

- 52% of angel deals returned zero or less than the capital invested
- 7% of the deals achieved 10X returns and represented 75% of all returns

  • Lemons rot faster than plums ripen (from L. Villalobos).   Portfolio companies went belly-up in three years while 10X returns took 5 years to mature.
  • Due diligence matters!  Angel investors that collectively spent less than 20 hours per deal on due diligence had total returns of only 1.1X (versus 2.6X overall for the study), while angel investors who collectively spent 40 hours or more per deal earned returns of 7.1X over the life of these investments.

 It’s a GREAT time to be an angel.  Find a group and jump in!

Bill Payne is the 2010 BNZ University of Auckland Business School Entrepreneur In Residence.