Metrics question

A new employee first encounters Glean through Slack, Teams, or the browser extension. Define the activation funnel from first touch to 'activated' for that user, including the specific events you would instrument and the activation threshold you would use. Then identify the top 2 product changes you would ship first to reduce time-to-value.

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Whether you can define a precise activation funnel with concrete instrumentation and translate the data into the highest-leverage product fixes.

Define the funnel stages explicitly: first touch (opens Glean in Slack, Teams, or the browser extension), first meaningful query, first useful result (clicked or acted on), and activated (a second session with a successful query within a defined window).
Instrument specific events: query submitted, result clicked, result marked helpful or unhelpful, and return-session timestamp, so each funnel stage has a concrete, measurable signal.
Set the activation threshold concretely, for example a second successful session within 7 days, since that captures habit formation better than a single successful first query.
Analyze the funnel to find the biggest drop-off point, for example many users query once but never return, which would point to a first-experience quality or awareness problem rather than a query-capability problem.
Identify the top two product changes based on where the data shows the steepest drop, for example a proactive first-query suggestion tailored to the user's role, and a lightweight nudge if no second session happens within the activation window.
Validate both proposed changes with a small experiment before a full rollout, measuring their effect on the specific funnel stage they target.

Names concrete, instrumentable events at each funnel stage rather than vague stage descriptions.
Sets a specific activation threshold, a second successful session within 7 days, giving the funnel a real, testable definition rather than a vague concept.
Ties the top two product changes directly to the biggest observed drop-off point in the funnel, rather than picking generic feature ideas.
Proposes validating changes with a small experiment first, showing rigor before a full rollout.

Defining funnel stages without specific, instrumentable events behind them.
No concrete activation threshold, leaving 'activated' as a vague concept.
Proposing product changes not clearly tied to the biggest drop-off point in the data.

How would you decide the right activation window, 7 days versus a shorter or longer period?
What would you do if the biggest drop-off is at first touch, before any query is even submitted?
How would you validate a proposed fix actually improved activation, not just first-touch volume?