How would you test a price increase?

##### Question You are a data scientist supporting a subscription B2C product: an AI video editing tool sold with a free trial and paid tiers. Product leadership is considering **raising prices**, for example increasing the monthly price of the Pro plan by 10-25%, and possibly changing what is included in each plan. Design the analysis and experimentation plan behind that decision. Specifically: 1. **Should we raise the price at all?** How would you analyze whether an increase is likely to be beneficial before you run anything? 2. **What exactly would you ship, and to whom?** The price change and/or the packaging change (list price, included credits, tier structure, annual vs monthly), and which plans, geos and customer segments it applies to. 3. **How would you estimate price sensitivity / elasticity** from the data you already have, and how much would you trust it? 4. **Design the experiment.** Unit of randomization, arms, population, duration, power, and specifically how you would handle **existing subscribers** versus new signups. 5. **What are your success metrics?** Primary, diagnostic and guardrail metrics, and the tradeoffs between them. 6. **Which segments matter and why does segmentation matter here?** For example new vs existing users, region, creator/professional vs casual, plan tier. 7. **What confounders, biases and risks would you watch for?** For example seasonality, competitor promotions, selection effects, plan switching and cannibalization, delayed churn. 8. **What would you do if you cannot fully randomize** the price (legal, app store, or fairness constraints)? 9. **How would you read the results and decide?** Stopping criteria, ramp plan, and the final rollout recommendation.