Instead of making one $2 million investment, make five $400k investments... If you're investing at a tenth the valuation, you only have to be a tenth as sure.
I don't agree with this. The investors have to be just as sure of the risk involved in each valuation as before in order to have the same expected value for the overall portfolio. However, investing $400K in 5 companies instead of $2M in a single company will reduce the variance of the return on investment.
I think it's a tradeoff for the VCs between variance in the portfolio and the amount of work involved in finding 5 times as many companies. Given the amount of funding they deal with, I can understand them leaning towards the companies looking for $2M rounds.
It seems obvious. But I've proposed to several VC firms that they set aside some money and designate one partner to make more, smaller bets, and they react as if I'd proposed the partners all get nose rings.
As I pointed out above, the partner has to be just as sure of each of the 5 bets as he would be of one. I'd react the same way if someone suggested I'd do better at my job if I worked 5 times as hard.
Suppose your threshold for investing in a startup is an n% confidence that they'll one day have a market cap of a billion dollars.
Suppose instead you split that investment between 10 companies at a tenth the valuation. How confident do you have to be that any given one will become a billion dollar company? You have the same percentage in all these companies that you would have had in the case of a single, big investment, so now you only need one of the 10 to succeed in order to get the same return.
There is a tradeoff though--accepting a smaller probability of success increases the variance, which deterministically drags down the returns. To see the effect, extend your argument to the extreme case of very very early valuations. You have $100k to invest, and you invest $1 each in 100k companies, each with a 1 in a billion chance of being a billion dollar company. Though your expected value is the same, at the end of the year, most likely you have no money left to invest for next year. The same effect holds in less extreme cases--if you increase your variance and lose 50% of your money, you can't just make it back by having a 50% upswing the next year--you need a 100% return to recoup it. So variance matters.
This makes complete sense if you assume that a company that only needs a few hundred thousand to "figure things out" has an equal chance of becoming a billion dollar company as a company that is looking for a multi-million dollar investment that "is already taking off." That's evident to us inside, but your own description of the companies is enough to make a VC understandably skittish.
I understand your overall point, and I agree with it, but I don't think this assumption was clear the first time I read through the article.
I wholeheartedly agree that convincing VCs of this isn't the way to go, and that others are going to make a killing by stepping in. My main point is only that the VCs aren't being irrational.
VC's likely believe lowering the dilligence and involvement level lowers the likelihood of success. If VC's believe, rightly or wrongly, the success rate drops from 10% to 5%, then by lowering selection criteria and involvement, they've lowered the rate of return by diversifying into a pool of lower quality. The key is whether those things really matter. In justifying their paychecks, however, they are vital.
> Suppose instead you split that investment between 10 companies at a tenth the valuation. How confident do you have to be that any given one will become a billion dollar company?
You'd have to be more confident in aggregate.
Let's say in the former case you have a 50% certainty with one company; and in the latter case you have 5% confidence with each of 10 companies.
The chance that NONE of them succeed to that degree is 0.95^10, or about 60%. So only a 40% chance someone will make it, vs. 50% for the former.
With smaller probabilities the difference is less, but guesstimating at billions is such a crapshoot. How about something more realistic?
How about doubling your money: Let's say you put $1 mil into the former company. If it doubles its worth, you've doubled your money. But in the $100k for each of 10 companies case, they all have to double their worth, or one has to grow 20x.
Let's say you think it's a 50% safe bet in the former case, but a whopping 90% chance in the smaller cases. The chance they ALL double up is 0.90^10, or about 35%. So you're probably losing money.
Sure, some might do 3x or 4x to make up for a couple of the flops, but you probably aren't going to let a 100k investment just die; you'll be sinking more into the money losers, so the successes have to do even better to make up for the bailouts.
I don't agree with this. The investors have to be just as sure of the risk involved in each valuation as before in order to have the same expected value for the overall portfolio. However, investing $400K in 5 companies instead of $2M in a single company will reduce the variance of the return on investment.
I think it's a tradeoff for the VCs between variance in the portfolio and the amount of work involved in finding 5 times as many companies. Given the amount of funding they deal with, I can understand them leaning towards the companies looking for $2M rounds.
It seems obvious. But I've proposed to several VC firms that they set aside some money and designate one partner to make more, smaller bets, and they react as if I'd proposed the partners all get nose rings.
As I pointed out above, the partner has to be just as sure of each of the 5 bets as he would be of one. I'd react the same way if someone suggested I'd do better at my job if I worked 5 times as hard.