Three layers. Not a formula. A way of looking at things.
See the thing.
Start with what is actually happening—not only the explanation already attached to it.
Watch the handoffs, exceptions, and workarounds. Listen to the people doing the work. Notice where the intended process and the actual day diverge.
The visible complaint is evidence. It is not yet the diagnosis.
Find the thing beneath the thing.
Ask what keeps recreating the problem, who carries its cost, and what makes the workaround necessary.
Look at incentives, information, responsibilities, and assumptions. A request for another feature may be pointing to a different problem underneath.
Understanding should change what you decide to build.
Build accordingly.
Build an answer people can try. Make the next action, responsibility, and limits clear.
Test it in use. Keep what helps, change what does not, and return to observation when the evidence challenges the original explanation.
An idea is not better until it works.
Observation
A request keeps falling between two teams.
Understanding
Responsibility changes hands. Ownership stays unclear.
Action
Make the handoff, the owner, and the next action visible—then test whether the work improves.
This is an illustration of the questions, not a universal diagnosis or a claim about a deployed product.
See how we think →