# 14A Following the Evidence to a Cause

Once a problem can be reproduced, I look for evidence that distinguishes possible causes instead of trying unrelated changes at random.
In the empty menu example, the browser showed a failed request, so the next useful question concerned the response from the server.
We compared the request with the expected endpoint and discovered that the application was using an incorrect configuration value.
That observation suggested a cause, but we still needed to check whether correcting the value restored the expected data and visible options.
After the correction, the request succeeded and the menu displayed the topics that were already stored in the database.
We then repeated the original steps with another lesson to see whether the same explanation fit more than one example.
This sequence connected the visible problem, the underlying request, the incorrect value, and the result of the specific change.
It also helped us avoid a misleading explanation in which refreshing the page appeared to be the entire solution.
A cause should explain the evidence and predict what will change when the relevant condition is corrected or removed.
The clearer that connection becomes, the easier it is to describe the fix and decide which related behavior still needs attention.