14.9 Explaining technical methods

Prepare a plain-language explanation for every model, metric, and projection assumption. A useful explanation covers:

1. what the method is trying to estimate;
2. which information it uses;
3. how performance was checked;
4. what the result means; and
5. what the method cannot establish.

For example: “The retention model estimates the probability that an eligible new member will remain active for twelve months using information available in the first sixty days. We tested it on members not used to fit the model and compared it with a majority baseline. It identifies a group for a pilot, but it does not establish why members leave or prove that contact will improve retention.”