What Is Worth Learning?

“We need more information” is usually true.

There is always more we could learn. More customers to talk to. More competitors to study. More workflows to understand. More assumptions to test.

The problem is that learning can expand almost without limit.

A better place to start is the decision.

What are we actually trying to decide?

Then work backward:

What uncertainty could change that decision?

What evidence would reduce that uncertainty?

Who or what could provide that evidence?

What is the cheapest credible way to get it?

The hard part is often not collecting the evidence. It is making the question precise enough.

Compare:

We need to understand how banks do this.

with:

We need to know whether this workflow is common.

and then:

We are about to make design choice X. It depends on assumption Y being true outside our current environment. Do we have enough evidence that Y generalizes?

Those are very different research problems.

The first might require months.

The last might require a few targeted conversations.

A vague decision creates a large information problem. A precise decision can make that problem much smaller.

There is a limit to this too. You can spend too long refining the question. At some point another hour of thinking teaches less than simply going out and getting evidence.

So the goal is not to eliminate uncertainty before acting.

It is to identify which uncertainty is actually carrying the decision, and learn enough about that one.

Once you know what would be useful to learn, there is a separate question: whether it is worth waiting to learn it.

That is a different problem.