You're reviewing an A/B test on a product change and the experiment readout shows a p-value for the primary metric. The team is debating what that number does and does not mean, especially after using CUPED to reduce variance.
How would you interpret a p-value in a product experiment?