You are reading an A/B test result for an Instagram Reels ranking change, and the primary metric is Instagram Save. The readout reports a p-value, and the team wants to know what that number actually means, especially since the analysis used CUPED.
How would you interpret a p-value in a product experiment?