Funnel optimization: ranking leaky steps by recoverable conversions

A product's conversion funnel has several steps, and users drop out at every one of them. Which step would you work on first, and how would you decide? ```hint The biggest leak isn't automatically the answer Before ranking steps, ask what fixing each one would actually return: how many users reach it, how much of its loss is realistically recoverable, and what the users you rescue go on to do. ``` ### Clarifying Questions - Which funnel is this, for example a sign-up-to-payment flow or an e-commerce checkout, and what counts as the final conversion? - Do I have user counts for each step, broken down by device and acquisition channel, and has any step's rate changed recently? - Is the goal more conversions soon, or more conversions that stay valuable, such as paying users who don't refund or churn? - Are all the steps ours to change, or do some depend on outside parties such as a payment provider or the marketing team? ### What a Strong Answer Covers - Checking the funnel data and separating natural drop-off (browsing, price checking) from fixable friction - A method that compares steps by expected additional conversions rather than by the raw size of each drop - How user intent at each step changes the value of the users a fix recovers - Effort, confidence, speed of learning and ownership as part of the ranking - A test plan with an end-to-end primary metric and guardrails on conversion quality ### Follow-up Questions - The largest absolute loss is at the very first step, which marketing owns. How do you handle that? - Your fix raises conversion at the payment step, but refunds rise too. What do you do? - How would your answer change if users take days or weeks to move between steps rather than one session?