Yeah, there's no deep data. But assuming it's roughly correct, iterating n quickly can be a very strong strategy. This is part of why the usual "early and often" advice is so effective.
The analysis is interesting, and I think it fits for low values of n (and even then, only if the VCs LOVE the team), but I think you might be looking too far ahead without doing the non-numerical analysis. If you failed seven times, do you think you'll get funding and have an 83% chance of success? I think your chance of getting funding, no matter how much the VCs love you, will plummet drastically after a certain value of n, and I'm guessing it's a lot lower than seven (but I have zero data to back that up).
Yes, it's an extremely naive first-order model. It assumes a series of independent "fair coin toss" events and gives the probability of at least one success using the given values.
Really, business success is not a random event, although there are many random factors which influence it. What this is doing is taking the "success rate" (which is whatever you want to make it be) and assuming that it works as a forward-looking probability.
In reality, the growth rate is likely to be higher for many people due to learning and the relatedness of successive ventures. There are also likely to be limiting factors which scale more rapidly than experience increases the probability of success - you're not likely to ever reach 83%.
But keeping in mind that it is a generalized approximation, the conclusion of the first-order model is still good.
* Start early, there are no ideal conditions.
* Evaluate often, if you don't do anything that can fail then you are likely over-invested in the single attempt.
"5. Better VCs Provide Better Deals: Venture capital firm experience is positively related to pre-money valuation. More experienced firms pay more for new ventures -- likely because they have higher success rates. "
7 comments
[ 3.9 ms ] story [ 23.7 ms ] threadp1 = 0.18
p2 = 0.20
Naively, p1 || p2 = 0.344
Cumulative probability after N attempts assuming p = 0.2 for any n > 1:
1: 0.180
2: 0.344
3: 0.475
4: 0.580
5: 0.664
6: 0.731
7: 0.785
8: 0.828
This also neglects things like mixed-experience teams (i.e. the probability data in the article is obviously very high level and overly simplistic).
Really, business success is not a random event, although there are many random factors which influence it. What this is doing is taking the "success rate" (which is whatever you want to make it be) and assuming that it works as a forward-looking probability.
In reality, the growth rate is likely to be higher for many people due to learning and the relatedness of successive ventures. There are also likely to be limiting factors which scale more rapidly than experience increases the probability of success - you're not likely to ever reach 83%.
But keeping in mind that it is a generalized approximation, the conclusion of the first-order model is still good.
* Start early, there are no ideal conditions.
* Evaluate often, if you don't do anything that can fail then you are likely over-invested in the single attempt.
Kind of tautology, no?