Metrics question

Figma launches a new AI feature, but very few eligible users reach the first moment of value. How would you identify the single biggest activation bottleneck across the funnel, from awareness to first successful output, and prioritize a portfolio of experiments to improve activation over the next quarter?

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

Tests funnel analysis skill to isolate the single biggest activation bottleneck and turn it into a prioritized quarter of experiments.

Break the funnel into explicit stages: aware of the feature, opened it, attempted a first action, and reached a successful first output.
Measure drop off rate at each stage rather than looking at overall conversion, since the biggest bottleneck could be anywhere in the chain.
Identify the stage with both the steepest drop off and the largest absolute user count, since that combination has the highest leverage to fix.
Form a hypothesis for why that specific stage is failing, for example unclear entry point placement or confusing first action instructions.
Design a portfolio of experiments targeted at that one bottleneck, varying entry point visibility, onboarding copy, and default settings, rather than scattering tests across the whole funnel.
Sequence experiments by expected impact and speed to run, running the fastest, highest confidence tests first.

Identifies the bottleneck using both drop off rate and absolute volume together, not rate alone, which correctly finds the highest leverage fix point.
Concentrates the experiment portfolio on the one identified bottleneck rather than spreading tests thinly across the entire funnel.
Forms a specific hypothesis for the bottleneck's cause before designing experiments, rather than testing blindly.

Looking at overall activation rate without breaking it into stages to find where the real bottleneck sits.
Running scattered experiments across the whole funnel instead of concentrating effort on the single highest leverage stage.

How would you decide when to stop iterating on this bottleneck and move to the next stage?
What would you do if awareness turns out to be the bottleneck rather than a usability issue inside the feature?