You ran an experiment on a product change and the analysis says the result is statistically significant. But the observed lift is tiny, and you are not sure it matters for the business.
How would you explain a statistically significant result that has very little business impact?