There is a ton of other "noise" as we all agree was caused by "visitors" to the VC world…but the data is really irrelevant to everybody but service providers.

The question is given that you have a company that is right for VC funding are there enough funds to provide funding and I think the answer to that is yes.

People are always going to say its hard to raise VC money and that's because it is…..however, I would say that most companies are better off not raising it and that's not a knock on VC's that's the way it should be.

I think Brad has a great example with Occipital and Josh Kopleman has written about it as well. I think frankly the hard part is having a company that is better off not raising VC have to compete with a company that raised VC money and now is spending it in a way that skews the market. (if you can steal their heat: i.e. leverage off of all the money they spend trying to artificially grow the market and outlast them it can work out well)

But back to the topic as I think Brad puts it well is "who gives a fuck?" and that would be the lawyers/accountants that publish/read/take stock in this crap.

If you are a VC or a entrepreneur why do you care? Are you going to stop investing or stop looking because of the data? Does it really affect you? No. It really only affects you if you suck a percentage of the money off in service fees.

10% feels about right to me and I agree that definitive data is frustrating difficult to find. Scott Shane, in his book "Fools’ Gold", says “Estimates using data from the "Entrepreneurship in the United States Assessment" indicate that the friends and family capital market is about $139 billion annually.” The best numbers I can find indicate that traditional VCs and angels each invest about $20 to 25 billion per year.

So the total seems to be somewhere in the $200 billion range, giving each of VCs and angels about 10% of the total. (I am still surprised by how big the Friends and Family component is.)

In the interest of keeping the debate going, I would like to respectively disagree with your point about using “longitudinal data – probably going back to the 1980s”. The VC industry has changed SO much since then. Back in the 1980s, the average VC principle was responsible for investing about $3 million – not per year, but total. Many of my angel friends manage portfolios bigger than that.

Back in the 1980s, the median VC investment in companies with M&A exits was around $5 million. Now it’s over $30 million.

Here are a couple of graphs to illustrate those metrics: http://www.angelblog.net/Venture_Capital_Firms_Ar…

The size of today’s traditional VC funds is the single biggest reason that the old VC model is so badly broken.

I like your term ‘visitors’ to describe a lot of the guys who built, and then failed to effectively manage, those huge VC funds created in the 1990s.

The economy desperately needs more early stage investors like you and Fred Wilson who will actually invest in start-ups. And more greybeard VCs like Alan Patricof who will say publicly that to produce a good return, today’s VC funds should be around $75 million in size. And of course, a lot more angel investors.

On the identity crisis, my comments may be more about my own ‘damage’. I co-founded a traditional VC fund in the early 2000s. About five years in, I realized that the ‘traditional VC model’ was broken – irreparably. It was painful to see all that work go into building something on a defective foundation. I don't want to identify with being a “VC” anymore because most people now associate that term with something that doesn’t work and needs to change.

I wish we could find a new term that followed in the size sequence of:
1.angel investor,
2.angel fund
3.<new name for a right-sized fund that works beneficially with entrepreneurs and produces an outstanding return for its stakeholders>

Actually I would argue that if your median investment size is over $5M you aren't really a VC.

Traditionally (i.e. before the late 90's) these type of investments were not even called VC deals. They were called Mezzanine Rounds and could be syndicated by some of the independent high tech investment banks which were swallowed up.

I've looked at this data before and agree its total crap….you can't even try to analyze it because its plain wrong. Small deals are totally unreported…..and this would be expected. The people who like this info/place stock in it are service providers. They care because the more deals/money gets raised the more revenue they can generate.

Just like "VC's" many of these guys have drank their own kool-aid and have gotten too big. Meaning you only like to do six figure engagements and that just doesn't work for a traditional startup.

Frankly if you're an entrepreneur the situation is binary: you get funded or not and I would agree as you list out your group if your idea is a right fit for VC funding there are resources out there.

For a good essay on the situation and data, you get a A. There were lots of places you could have gotten off track but did not. Good.

Generally to the readers, I would advise: When using data and statistics, be clear about what you are trying to conclude. Instead, too often people just toss out the mean, median, other percentiles, confidence intervals, etc. as if the point were to rise from sea, fall from the sky, etc. which it rarely does.

In particular, yes, if have a random variable, then necessarily it has a distribution. My advice: Mostly f'get about the distribution. Certainly don't ask if it's Gaussian or log-normal. In particular, just talking about the distribution will make no useful truth rise from the sea.

Yes, as in S. Gould, what you are more likely interested in is the conditional distribution given what additional, relevant data you have, and that can be much different. E.g., what is the conditional distribution of first round funding amounts given that the project was in IT instead of biotech? The distribution, expectation, percentiles are not something 'immutable' and typically change a lot given more information. E.g., in 'Wall Street', the conditional expectation of what the steel company was worth changed a lot given that Sir Raider was flying his Gulfstream to PA.

Here are some cases where can talk about a distribution:

(1) An 'arrival process' with 'stationary, independent' increments will be a Poisson process with independent, identically distributed (iid) times between arrivals with exponential distribution. Nice. Qualitative assumptions, often can check just intuitively, with precise quantitative consequences. See Erhan Cinlar's 'Introduction'.

Possible application: What is the probability a small package cargo airplane will be too heavy? Assume number of packages does not affect the distribution of the package weights; use Poisson for the number and historical data for the weights.

(2) Under mild assumptions, the average of n iid random variables converges to the mean with probability 1 — strong law of large numbers. E.g., in betting, in the long run what you get per play is the mean. Nicest proof from the martingale convergence theorem (Leo Breiman, 'Probability'). Yes, should realize that the rate of convergence does have to do with the population of 'black swans'.

(3) Under mild assumptions, the distribution of the sum of n iid random variables divided by the square root of n converges to a Gaussian distribution. Central limit theorem (check Lindeberg-Feller). Black swans are again relevant but notice the Berry-Essen bound.

For answering some questions, (1)-(3) can be useful.

For the VC data, some candidate questions:

(1) Is the whole world of finance drying up with growth stopping and all of us going to hell? Not yet.

(2) If an entrepreneur has a good project, can it still get funded? Likely.

Generally to the readers, I would advise: When using data and statistics, be clear about what you are trying to conclude. Instead, too often people just toss out the mean, median, other percentiles, confidence intervals, etc. as if the point were to rise from sea, fall from the sky, etc. which it rarely does.

In particular, yes, if have a random variable, then necessarily it has a distribution. My advice: Mostly f'get about the distribution. Certainly don't ask if it's Gaussian or log-normal. In particular, just talking about the distribution will make no useful truth rise from the sea.

Yes, as in S. Gould, what you are more likely interested in is the conditional distribution given what additional, relevant data you have, and that can be much different. E.g., what is the conditional distribution of first round funding amounts given that the project was in IT instead of biotech? The distribution, expectation, percentiles are not something 'immutable' and typically change a lot given more information. E.g., in 'Wall Street', the conditional expectation of what the steel company was worth changed a lot given that Sir Raider was flying his Gulfstream to PA.

Here are some cases where can talk about a distribution:

(1) An 'arrival process' with 'stationary, independent' increments will be a Poisson process with independent, identically distributed (iid) times between arrivals with exponential distribution. Nice. Qualitative assumptions, often can check just intuitively, with precise quantitative consequences. See Erhan Cinlar's 'Introduction'.

Possible application: What is the probability a small package cargo airplane will be too heavy? Assume number of packages does not affect the distribution of the package weights; use Poisson for the number and historical data for the weights.

(2) Under mild assumptions, the average of n iid random variables converges to the mean with probability 1 — strong law of large numbers. E.g., in betting, in the long run what you get is the mean. Nicest proof from the Martingale convergence theorem (Leo Breiman, 'Probability'). Yes, should realize that the rate of convergence does have to do with the population of 'black swans'.

(3) Under mild assumptions, the distribution of the sum of n iid random variables divided by the square root of n converges to a Gaussian distribution. Central limit theorem (check Lindeberg-Feller). Black swans are again relevant but notice the Berry-Essen bound.

For answering some questions, (1)-(3) can be useful.

For the VC data, some candidate questions:

(1) Is the whole world of finance drying up with growth stopping and all of us going to hell? Not yet.

(2) If an entrepreneur has a good project, can it still get funded? Likely.

Taking skewed, non-normal (& possibly kurtotic) data and torturing it until it fit a normal distribution destroys information & when done badly adds bias & skew. Worse… It's all too tempting to bend the data until it fits our preconceived notions.

Anyway, thanks!

Re: The data is that angels invest in about 27x more startup and seed deals than traditional VCs: http://www.angelblog.net/Angels_Finance_27_Times_… But, isn’t overall VC investment less than 10% of the total investment in private companies?  If so, this data would simply be confirming the ratio of activity.  I don’t have the data (and don’t feel like hunting around for data that shows this), but I’ve heard many times – from many different people (academics, angels, angel group leadership, VCs, entrepreneurs, other) that VCs invest in “less than 10% of the private companies funded each year.”  If this is true, isn’t your assertion about VCs only doing 4% of the angel investments in the measurement period simply directionally confirmatory data? 

Re: the assertion that is going around that VCs have “effectively stopped investing in seed stage ($500k and less) and startup-stage ($2m and less), I think it’s important to look at longitudinal data – probably going back to the 1980s.  One would need to normalize this data by something for the period from 1998 to 2001 (the Internet bubble) when there was a huge spike in investment in the VC category.  I’ve never seen data that I would consider “good” on this so I only have my own experience, but the tempo (# of these investments / year) doesn’t seem to be declining precipitously from my perspective.  Yes – it bounced around from year to year, but the overall trend doesn’t feel like an curve that is asymptotically approaching zero.

Also, in both cases, don’t forget to try to correct for survivorship bias.  Over the last 15 years there have been a lot of “visitors” in the VC category – people (or firms) who are VCs for less than a decade.  They entered in 1998 – 2001, were unsuccessful as VCs (or funds), and started exiting around 2005.  There’s a second wave of this going on right now with funds raised in the 2000 – 2003 time frame, that have not been successful (and hence not able to raise a second fund), and are now winding down their funds.  Many of these both of these categories made many seed / startup-stage investments early in their fund life (especially since they were smaller funds so they positioned themselves as seed investors), aggressively invested for 3 to 5 years, and then abruptly stopped investing because their five year investment period (where they could make new investments) was over.  In these cases, they haven’t made a new investment since 2006!  However, by 2011, they won’t be counted in any numbers any more since they’ll be out of the industry.

Re: Identity Crisis: I’m very comfortable with my identity!  Having been an entrepreneur, and angel investor, and a VC, I’m most definitely a VC.  I just happen to be an “early stage VC”.