Find leading indicators for long-term churn

WHAT THIS TESTS: This question tests your ability to apply a structured, scientific method to a vague business problem. The interviewer is not looking for a single SQL query. They are evaluating your product sense, your data-driven mindset, and your ability to move from analysis to action. Can you break down a lagging indicator (12-month renewal) into tangible, measurable, and influenceable leading indicators from early user behavior?

A GOOD ANSWER COVERS: A strong answer outlines a clear, four-step process. First, define the cohorts by separating users into two distinct groups: those who renewed at 12 months and those who did not. Second, hypothesize which user actions within the first 30 days are critical for a user to experience the product's 'aha!' moment. This requires product intuition. Third, perform a comparative analysis, measuring the frequency of these hypothesized behaviors between the 'retained' and 'churned' cohorts. This is where you'd use cohort analysis to find statistically significant differences. Fourth, propose an actionable leading indicator and a plan to influence it, such as running an A/B test on a new onboarding flow designed to encourage that specific behavior.

COMMON WRONG ANSWERS: A major red flag is jumping immediately to complex solutions like, "I'd build a machine learning model to predict churn." This is premature for an initial exploratory analysis and signals a focus on tooling over strategy. Another weak answer is a vague plan like, "I'd look at the data for patterns." This lacks the structured approach expected of a senior role. Finally, focusing only on high-level metrics like Daily Active Users (DAU) is insufficient; a good answer digs into specific, value-creating actions.

LIKELY FOLLOW-UPS: Expect follow-ups like: "What if you find a correlation but aren't sure about causation?" (The answer is to propose an experiment, like an A/B test, to establish causality). "How would you define an 'active' user for this product?" (The answer should be tied to the product's core value proposition, not a generic definition). "What specific tools or queries would you use?" (Mention SQL for data prep and a product analytics tool for behavioral/cohort analysis).

ONE CONCRETE EXAMPLE: After analysis, we find that users who invite at least one teammate within their first 30 days have a 65% renewal rate, compared to a 20% renewal rate for those who do not. This makes 'team invite sent' a strong candidate for a leading indicator. The actionable next step is not just to monitor this metric, but to design an experiment (e.g., a new in-app prompt or a simplified invite UI) to increase the percentage of new users who send an invite, and then measure if that change positively impacts the 12-month renewal rate.