How Do Venture Capitalists Make Decisions?

[hoodoof]

If it's already a success in a global market, with a complete team, a fully complete product, clearly making plenty of money with tens of thousands of paying customers (we'd really prefer to see millions) and all risk removed then WE INVEST!

[hoodoof]

AND we want a board seat, 45% of the company and a bit of damn gratitude.

And a new CEO, the founders, well......who are these inexperienced bozos?

[hugs]

Lots of entrepreneurs would save so much time and headaches if they realized how right you are.

[ulfw]

Yep exactly. Great to invest when there is really no need for it at all.

[hodder]

Given the dismal returns on the median venture capital firm, perhaps the criteria identified in the paper should be a guide to how NOT to make investment decisions.

[djyaz1200]

Agree! I'm far more interested in how they con LPs into paying them to lose money. Why should entrepreneurs do all the work to build companies, create jobs and solve problems for customers when we could just get paid to lose money for pension funds?!

[gohrt]

median is wrong metric for an asymmetric distribution.

[lowglow]

Surveying might yield skewed results. Perception is reality to outside players in venture capital. I actually wrote an article about what I (a Hacker) observed sitting on the other side of the table during my time working at a Venture Capital firm.

Check it (forgive the slightly tongue-in-cheek writing): http://www.techendo.com/posts/what-venture-capital-companies...

[hammerzeit]

This is a really interesting question to ask, but survey-based responses only tells us how venture capitalists think they make decisions. Obviously this is a secretive industry but I think the far more interesting question to answer is around revealed preferences rather than self-assessment.

[known]

Hope someone explains in Feynman Technique http://qz.com/849256/how-to-master-a-new-subject/

[hugs]

  function decide(marketSize) {
    if (Math.random() < .01) {
      return "term sheet"
    } else {
      return undefined
    }
  }

[danieltillett]

Almost right :)

  function decide(connected) {
    if (connected || Math.random() < .01) {
      return "term sheet"
    } else {
      return "keep us updated on your progress"
    }
  }

[mathattack]

This is very interesting research. I'm interested in anything that goes beyond Survey data to what actually happened. (VCs may say founding teams matter most, but I also say that I eat less than 2000 calories a day - memory is suspect) I'm also interested in how this would look weighted on firm size or returns. There are lots of small VCs out there who don't make money. How do the VCs who funded Google/Facebook/Apple/Oracle invest?

There's a ton of academic research on public equity investment decision theory. It's great to see private market investment get some focus.

[graycat]

On page 2, the paper has

> In fact, Kaplan and Stromberg (2001) and Gompers and Lerner (2001) argue that VCs are particularly successful at solving an important problem in market economies|connecting entrepreneurs with good ideas (but no money) with investors who have money (but no ideas).

IMHO, for information technology (IT) venture capitalists (VCs), this statement about "ideas" is mostly wrong. One reason the statement is wrong is that VCs will rarely even look in any detail at an idea.

In contrast the US NSF and DoD will look very carefully at ideas, e.g., GPS, stealth, measuring the 3 K background radiation. So, will Ph.D. dissertation committees, reviewers at peer reviewed journals of original research, and more. IT VCs won't do such things.

IMHO what IT VCs look at is current traction, that is, users or revenue, and want that traction to be significant and growing rapidly. Then if nothing else is wrong -- team, competition, scalability, etc. -- an IT VC can get very interested.

So, the VCs want the idea already implemented in hard/software (usually software) and in the market and in front of users/customers.

In the world of VCs, the idea is not something from research, that could be in a peer-reviewed papers, etc. but is just a short description of what the business looks like externally to a casual observer, the common man in the street. That there could be anything from a research idea as the crucial core of the business, crucial for getting the traction, being defensible and scalable, etc., is just ignored.

So, suppose some IT founding team has the coveted traction. If they have lots of users, then the team should be able to run ads and get significant revenue. If they have lots of customers, then they should also have significant revenue.

With that revenue, there will be some serious question if the team should accept equity funding, that is, accept the terms, a Delaware C-corporation, the BoD, reporting to the BoD that can fire team members, etc. A C-corp and a BoD bring a lot of overhead.

Really, the example of the romantic match making service Plenty of Fish (PoF) starts to look more important as a example for IT startups in the future. PoF was long just one guy, two old Dell servers, revenue just from ads and the ads just from Google, and $10 million a year in revenue. As in

http://techcrunch.com/2015/07/14/match-group-buys-plentyoffi...

on about July 14, 2015, the solo founder Markus Frind sold out for $575 million.

If a solo founder has a good idea that needs mostly only software, then there is a good chance they can just write the software, bring it to market, get traction, and have revenue enough for rapid organic (that is, funded by earnings from revenue) growth.

That is, a solo founder who invented the idea can keep costs, time on communications, etc. low, write the software, go to market, and get the traction. That day is the first a VC wants to hear from that founder, and it is likely the last day the founder would be willing to accept a check, term sheet, etc. from a VC.

Net, with the VC rules, by the time the VC is willing to invest, the solo founder is beyond willing to accept the check.

Of course, if there are several founders, some high burn rate, maxed out credit cards, each of the founder with a pregnant wife, etc., then the VC's check might be more welcome.

But we need to understand: All across the US, entrepreneurs start and grow businesses -- pizza shops, auto body repair, dentist's office, etc. -- without VC investing. Then, the big difference for an IT startup is that some software and current computing, the Internet, etc. can let an entrepreneur make money much faster than a pizza shop. E.g., suppose the founder's business is a new Web site, and a lot of people like to connect. Suppose the site sends 10 Web pages a second with each page with five ads. Suppose get paid (from a report from Mary Meeker at VC firm KPCB) $2 per 1000 ads displayed. Then the monthly revenue would be

     10 * 5 * 2 * 3600 * 24 * 30 / 1000 =   
     259,200

dollars. Heck, sending even just 1 page per second would yield $25,920 a month. Then one founder with $25,920 a month in revenue growing rapidly is just the traction VCs want, but, with that traction, why should the founder want to accept the VC's check?

[taneq]

> Net, with the VC rules, by the time the VC is willing to invest, the solo founder is beyond willing to accept the check.

This is what I don't understand. So often someone is in this position, dips their toe in the VC pool, and a VC says "here, have two million dollars but give me control." Then suddenly their formerly highly-profitable startup has to have 50 staff, a big office building with board rooms and lots of shiny glass, and before they know it they're living off successive rounds of funding then filing for bankruptcy, because a $1mil/quarter 1-man company can't magically produce $50mil/quarter from the same banner ads just by hiring 49 more people.

Why would you not just tell the VC "no thanks, I'm doing fine without interference"? Your company is real at that stage. The dollar signs being waved in your eyes by the VC are not.

[Retra]

I'm a very stupid and naive person when it comes to this kind of thing, but if I were such a founder, I'd be a little worried that such a VC would invest in a direct competitor to me if I didn't take their check. Of course, they might do that anyway (and other's surely would,) but at least I can imagine competing better if I had more funding.

[kriro]

It's interesting sure but I feel like a different method is needed. I think qualitative research would be more valuable at this stage. A large field study would be ideal. Embed researchers with VCs possibly also with startups who try to raise for an interesting overall picture. Obviously not the easiest thing to pull off.

[dpandey]

This paper is simply a comprehensive survey of VC opinion broken down statistically, rather than a genuine study into how VCs work.

In other words, the singular focus on survey data without a sharp focus on correct interpretation has killed the spirit of digging deep and understanding the VC setup, with the result that it's not very actionable or insightful.

If the findings were perhaps informed by cross referencing with a survey of others in the ecosystem who aren't VCs, the authors would probably have been more skeptical and possibly better educated on reality. It'd have been extremely valuable to go to entrepreneurs and ask.

Reading one of PG's essays provides a far more useful perspective.

The challenge with a generic survey such as this is that while it gives an (really long) introduction into how VC works, it misses all the subtleties. A better researcher would have looked into that, and turned this into a legendary paper. Why is that important? Because as a VC, decisions are the most important thing you make, and subtleties play a huge role in that. If you don't dig deep, then you're just documenting interviews rather than understanding and interpreting reality correctly.

What the authors should actually have focussed on is:

How do VCs really make decisions?

[myf01d]

Like any other kind of thieves, I guess.

[bayesian_horse]

Optimistically, of course.

[dilemma]

feels

[CalChris]

It's behind a paywall. Publicly accessible version:

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2801385

Actually, it's a pretty good read if long (89 pages). Kinda lite on angels but it covers syndication.

[sctb]

Thanks, we've updated the link from http://www.nber.org/papers/w22587.