There are studies out there for success rates of VC-funded companies. I read one that had a success rate of 18% for first-time founders, 20% for founders that had previously failed, and 33% for founders that had previously succeeded. Don't have time to dig it up now, unfortunately - it was about the gains from serial entrepreneurship, and I think it was posted on news.YC.
It's much harder - possibly impossible - to measure the success rate of bootstrapped ventures, because the boundaries are much fuzzier. For example, my current venture started with 5 founders (none of whom are still with it) and is now just me. We've launched 3 websites, all failures, and have one still under development. If the new website succeeds, should this count as 5 failures, one per founder that left? 1 failure, treating the original founding team as a failure and me as a success? 0 failures, since none of the other 5 founders quit their day jobs, and so it was just a hobby for them? 1 failure, since one of the other founders committed significantly to this and did a lot of work for it (even while still holding his day job), yet the other 4 didn't really care? 3 failures, one per failed idea? 1 failure, since two of the failures were launched while we were all still at our day jobs? 0 failures, since it all worked out in the end? 2 failures, since the hypothetical success builds upon one of the failures? 5 failures, since there were a couple other hobby projects that weren't intended as startups but might've become so had there been interest?
With VC, there's a clear delineating line between startups. You just go by the corporate entity, which must exist for them to take investment. But bootstrapped startups frequently don't incorporate until it looks like they'll be successful, and often success comes from accidents.
And by many of the definitions of bootstrap failure, the big VC-funded startups are failures too. My first employer (VC-backed) went through 4 business plans in the year I was there. Paypal did too, as did Flickr. Should those be counted as 3 failures and 1 success?
I'm skeptical about any sort of statistical data anyway, because (as with any economic data) the act of publishing that data changes the data itself. If a study comes out showing that you should not take VC money, there will be a flood of entrepreneurs into those markets that can be serviced without VC funds. That'll necessarily drive down success rates, as the market becomes oversaturated, competition drives down margins, and bootstrappers find it unprofitable to continue. Similarly, if a study comes out showing that you should take VC money, there'll be a flood of entrepreneurs into the markets that are considered "hot" by VCs, like what happened in social-networking/bookmarking in 04/05. That'll drive down success rates, companies will fold or be acquired for peanuts, and the next study will say you shouldn't take VC money.
You make a good point that I haven't considered. I would define "success" or "failure" as the ultimate outcome of the efforts of the group of founders who worked together on one or more products, regardless any part/full time jobs they had. Having one successful product, no matter how many failed ones, equals success. Having no successful products equals failure. That's just my definition -- feel free to pick a different one if you like it better.
One should be skeptical of any statistical study (didn't you know that 60% of statistical studies are flawed? :) ) but I think it would nonetheless be interesting to analyze a larger data set of startup outcomes to answer the VC funding question.
We've launched 3 websites, all failures, and have one still under development. If the new website succeeds...
If the new site succeeds the whole thing is a success. Who cares about the failures? Certainly not you.
Once you succeed the narrative will immediately fall into place. Your earlier "failures" will turn out to have just been stumbling blocks, correctable mistakes, and/or learning experiences along the road to success.
Which just makes your main point stronger: You can't rationally measure the "success" or "failure" rate of "ideas" until you define "success", "failure", and "idea" -- all of which are arbitrary and vary from company to company, from founder to founder, and from time to time.
It's much harder - possibly impossible - to measure the success rate of bootstrapped ventures, because the boundaries are much fuzzier. For example, my current venture started with 5 founders (none of whom are still with it) and is now just me. We've launched 3 websites, all failures, and have one still under development. If the new website succeeds, should this count as 5 failures, one per founder that left? 1 failure, treating the original founding team as a failure and me as a success? 0 failures, since none of the other 5 founders quit their day jobs, and so it was just a hobby for them? 1 failure, since one of the other founders committed significantly to this and did a lot of work for it (even while still holding his day job), yet the other 4 didn't really care? 3 failures, one per failed idea? 1 failure, since two of the failures were launched while we were all still at our day jobs? 0 failures, since it all worked out in the end? 2 failures, since the hypothetical success builds upon one of the failures? 5 failures, since there were a couple other hobby projects that weren't intended as startups but might've become so had there been interest?
With VC, there's a clear delineating line between startups. You just go by the corporate entity, which must exist for them to take investment. But bootstrapped startups frequently don't incorporate until it looks like they'll be successful, and often success comes from accidents.
And by many of the definitions of bootstrap failure, the big VC-funded startups are failures too. My first employer (VC-backed) went through 4 business plans in the year I was there. Paypal did too, as did Flickr. Should those be counted as 3 failures and 1 success?
I'm skeptical about any sort of statistical data anyway, because (as with any economic data) the act of publishing that data changes the data itself. If a study comes out showing that you should not take VC money, there will be a flood of entrepreneurs into those markets that can be serviced without VC funds. That'll necessarily drive down success rates, as the market becomes oversaturated, competition drives down margins, and bootstrappers find it unprofitable to continue. Similarly, if a study comes out showing that you should take VC money, there'll be a flood of entrepreneurs into the markets that are considered "hot" by VCs, like what happened in social-networking/bookmarking in 04/05. That'll drive down success rates, companies will fold or be acquired for peanuts, and the next study will say you shouldn't take VC money.