You run an A/B test on a customer-facing product change and see a positive result in the experiment population. Before acting on it, you want to understand whether the measured effect would hold for the broader customer base, including segments that may have behaved differently or been underrepresented during the test.
How would you explain why an experiment result might not generalize to all customers? What would you look at in the experiment design and analysis to assess whether the observed lift is likely to hold more broadly?