Start by making sure that you understand what the goal is. Frame the goal in terms of one of Dave McClure’s AARRR metrics.

Once you know what the goal is behind the idea (it should also obviously be tied to revenue), figure out what the most risky assumptions are. What are the parts where you’re not sure if people will click a button, or if they will not go through with something?

Once you’ve identified the riskiest, and most high-value assumptions, design a feature specifically to test those assumptions. Set success and failure criteria. It should take into account the fact that you need an interception point, i.e. a way to get customers in front of the feature.

After you’ve thought of the feature, cut it down so that it only tests the most important assumption. Don’t build anything extra.

Then deploy it and see if your assumption is validated. Repeat.