Start with the question
Before running another experiment, write down the mechanism you expect to observe. What behaviour should create the signal? Under which conditions would you expect it to disappear? A clear hypothesis makes both the result and its limits easier to interpret.
Change the conditions
Look at how the idea behaves when assumptions change. Compare different periods, inspect transaction costs, and keep an untouched evaluation period. Record the choices you made along the way so another researcher can understand the experiment.
Keep the evidence together
The useful output is more than a curve. Keep the data window, configuration, observations, and unresolved questions beside the result. That context is what makes the next experiment informed rather than accidental.
A result starts the conversation. A process helps you ask the next question.