Charlie Songhurst

studied politics first with economics at Oxford, ended up very briefly at McKinsey, then at Microsoft, and then sort of bumbled my way into becoming an investor and do a lot of angel investing.

So I think I've done about 500 angel investments.

I think the actual number is 483 and about 300 in the portfolio at the moment.

There's so many memories working with amazing people.

The hostile acquisition of Yahoo as an attempt was just unbelievably intellectually fascinating.

Trying to buy a company in a hostile acquisition for I think it's $47 billion and just the drama and the soap opera.

realizing how much path dependency matters on big deals, how much a casual comment misinterpreted by one side or the other actually changes the outcome.

It's remarkable how little system theory there is in the sense of if you ran experiments again and again, I think you would get a lot of different results in M&A activity.

Whereas in things like product usage, good products and bad products would probably be the same in each experiment.

In some ways, they're too tied together.

So I think the dominant sort of failure mode for startups is the same at each different stage.

So sort of pre-seed to seed, you basically have a failure to achieve labor productivity, which is really just a polite way of saying the team doesn't come together, doesn't gel and produce good output.

Usually a team that just produces good work will generate enough sort of kinetic energy to get continuing funding.

Once you're sort of going from seed to series A, it's another single cause, which is failure to get product market fit.

You're basically on the search for demand curve, and you either find one or you don't find one.

And this is where you've got the highest element of pure chance in a startup.

It sort of always feels akin to sort of gold prospecting in the California gold rush, which is you can be good or you can be bad.

You can do the right thing.

You can do the wrong thing.

But there's some irreducible amount of chance.