Horizontal-vertical thinking for startups

In startups, it can be very difficult to progress quickly when you're unsure about the optimal strategy. On the other hand, it's easy to execute a clear strategy quickly, because you simply have to follow the instruction manual. So how do we move fast when there is no instruction manual?

One parallel exists in reinforcement learning (a type of machine learning), where there is a trade-off between exploration and exploitation. Similarly, Sam Altman talks about knowing the difference between testing mode and knowing when things are working and you just need to scale.

The idea of horizontal thinking is that, in the exploration (or test) phase, you want to "go horizontal". That is, you want to give up the idea of making "forward" progress in the company and instead focus on rapid experimentation and research to obtain the information required to go forward in a certain direction, which then becomes "going vertical."

A startup is unsure whether to tackle market A or market B. They make a decision, given the information they have, to tackle market A.

Leave a Comment