Thread by @PavelASamsonov

One-way door decision: irrevocable. Do copious research before you do this, and have a plan B for the plan B.

Two-way door decision: reversible. It's ok to just do it, and learn whether you were right after the fact.

But there is a third kind of decision this framework misses.

I don't want to call them "three-way door" decisions so I'll settle for quantum decisions - decisions where the very act of doing *anything* changes the calculus.

You can make the decision, but results are uncertain. You can undo it/redo it, but the outcome will be different.

Most "two-way door" decisions are actually this kind of decision. I can go outside to see if it's raining, but I can never return to the initial conditions of not having gone out. I can return an ugly suit, but people will remember me having worn it.

This is the uneven ground on which "move fast and break things" falls apart. No matter how cheap it is to launch, you'll never be able to take back the product's first impression on your user base, especially if it's harmful.

In this way, decisions become about bandwidth.

The "two-way door" attitude is a framework for experiments. For an experiment to be meaningful, you must eliminate confounding factors.

So experiments become more about bandwidth: how much can we try before we change the initial conditions too much, and pollute the space?

And of course, there are layers. What if you go through one door, go back, then go through the other door - and it turns out that even though the first door was not as good as you'd hoped, the second door is worse? Too bad.

Parallel experiments permit revolving-door decisions.