The Anatomy Of A Pass, A Quantitative Analysis On Why A VC Passes | Hacker News Reader

The Anatomy Of A Pass, A Quantitative Analysis On Why A VC Passes

techcrunch.com

37 pointsssclafani15 comments

Meta-rule: Why VCs say yes or no is less complicated than they tell you, and possibly less complicated than they themselves believe.

1. If you seem formidable and have rapid growth of a sort that seems likely to continue, few VCs will turn you down.

2. If you don't seem formidable, few VCs will fund you.

3. If you seem formidable but don't yet have growth, whether or not you get funded depends on how convincing you are.

I know you're pg and all, but this comment seems almost content-free without defining "formidable". What does that mean?

I'll take a stab at it (though I recognize the entire purpose behind this being a "meta" rule is not to deconstruct).

It's a combination of intellectual capacity, previous track record of success (in something great - not necessarily startups, though that's a bonus), absolute-and-utter commitment to the project, incredible confidence in what you're doing, a team that's fully formed, awareness of all relevant landscapes (technological, regulatory, competitive, etc..) the ability to answer every question without hesitation, insight into where you are strong, and where you are weak, but, perhaps most importantly, the ability to project a sense that regardless of what obstacles are placed in your way (and there will be many), you will be successful.

Those teams always seem to come out ahead in the end.

The way I read it was that to "seem formidable" is first and foremost a subjective assessment. You have to impress the VC's as to your ability to succeed, in a short period of time. It's a smell test. I could be wrong. If I am, maybe Pg will correct me ;-)

"I know you're pg and all" -- ballsy, accurate and funny. Respect! :)

"Formidable" is entirely in the eyes of the beholder. This means what's formidable to one investor is not to another, for a wide range of reasons - market focus, experience with investments in the area of the venture's focus, and the intelligence and savvy of investors. In other words, it is entirely subjective and not particularly meaningful.

In a prior post on a related subject, I related that most VCs dramatically underperform the market and are not worth the investment [1]. This would imply that most VCs are not particularly good at determining the quality of their investments. I would argue that "formidable" is meaningless in this context when most VCs do such a poor job of distinguishing "winners" from "losers".

[1] http://www.kauffman.org/newsroom/institutional-limited-partn...

https://www.google.com/search?q=define+formidable

Mathematically speaking, this analysis is almost worthless. The inputs are not quantative in any way - he assigns values on a 1 to 5 scale, based on how he feels company to company rather than taking some representative metrics.

Beyond that, he has for some reason assumed that the error here is normally distributed (which it plainly is not as data is discrete) and applied OLS techniques to find his coefficients.

There is no discussion on statistical significance, on model construction, or on any subject which would allow this to be called quantative analysis. Even his r^2, the fallback number for people who don't understand stats to demonstrate how 'good' their model is, is woefully bad.

There is no insight here that would not have been given by expressing the same sentiments about pitch/investor fit without recourse to pseudo-statistics.

I think they are meaningful. Basically what he is doing is identifying subjective impressions and approximately how they affected the final result.

This is pretty helpful because what it shows to me is that if you are pitching to him, you want to emphasize team and market potential, and then touch on product and traction.

Where it would be less helpful is in trying to use it to measure performance of the VC's funds since these things are subjectively rather than objectively measured.

I have made money on much worse rsq. That said, I agree. There is no Team score, it is really BlueRun's Estimator Of Team Score, which has its own error, etc etc etc. 200 samples is not enough to back out 20 variables.

A better model might be constructed around a decision tree mechanism but I have less familiarity with that.

Finally it probably doesn't include statistics around which companies actually got money from some VC not just this VC.

I suspect this is MBA statistics -- only statistics that can be done in Excel. Anyone want to find his bio?

The other thing to note "all funded pitches were mobile centric"

This regression produces a score of 4.3 for a company that scores a perfect "5" in all 4 categories. This is a high score, but it's still closer to "pass, but with close monitor" than to "agree to invest & offer term sheet."

I'd like to see a similar analysis, but with a "fit" variable. The article suggests that fit (especially with regards to mobile vs non-mobile) was a critical factor in the funding decision, but it wasn't included in the analysis.

The idea that a quantitative analysis can be performed as a means to determine the acceptability for an investment for what are fundamentally and primarily subjective decisions made on discrete investments with no correlation is ludicrous. At its core, VC investments are made as a combination of the experience of the investors in a particular market, the ability for the venture to convince the investors they are worth the risk, and non-quantitative factors that are highly dependent on inter-personal relationships.

Long story short, if they like you and your market and believe you, they'll give you money. If not, they won't. That's pretty much the extent of any sort of "quantitative" analysis you can do here.

I suspect the OP wanted to imply there was a quantitative element to their analysis so that they can assert that they don't make "shoot from the hip" decisions and it's all based on some sort of model. But it's not. There's no quantitative basis here despite their assertions otherwise. There's more analysis in your typical neighborhood investment club than what happens in a typical (poorly performing) VC.

I actually thought this was a very interesting and useful article. It provides an insight into the mind of the VC as to he finds impressive, and I think this is helpful in figuring how to pitch.

There's something that is to be said about knowing one's audience and the measure of utility of an article like this is how well it helps one to know the audience.

With an r^2 of 0,5 this model only explains 50% of the variations in the dependent variable. In other words, this model is not that reliable.

Reply on news.ycombinator.com