Explore, Expand, Extract

Kent Beck discovered this pattern while trying to figure out why Facebook is successful despite having terrible engineering practices.

He goes on to try Coaching people internally. So he starts a coaching program called Good to Great, for 12 sessions, modeled off conversations with Ward Cunningham to see if Engineers could grow faster.

Turns out these were really useful to developers, but Facebook is metrics driven so the value wasn't being realized. But as time went on, Kent was able to show that students who did that were twice as likely to get promoted than their peers, which was a metric they did care about.

There's a big difference between payouts, convex and concave. This concept is adopted from a talk to Nassim Nicholas Taleb on his article To Prevail in an Uncertain World, Get Convex

If you make a small investment and most of the time you loose everything but every once and a while you get a huge payoff, this is called convex. If you're in a convex world, you want to make lots of these bets.

Contrast that with a world where you have large investments. Most of the time you have a small payoff, and some of time you loose a while pile of money.

In a concave world, you want to move the payoff up just a little bit because it makes a huge payoff and reduce the chances of a real failure.

Exploration is like you try a few things, but suddenly you do something similar to everything else but now you've found a growth loop-- the goal of Exploration is finding this loop where the more you grow the easier it is to grow more. It's the search for a new growth loop like that.

This isn't a rational process, so it's hard to find where the next loop will be. So you have to try joining different ideas together, like throwing messages away after 24 hours in Snapchat. For some reason that took off.