Metrics question

How would you measure the success of Uber Instant Pay?

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 building a metric tree for a fintech feature bolted onto a driver marketplace, balancing driver behavior with platform cost.

Confirm the goal: does Instant Pay aim to boost driver retention and peak-hour supply, or mainly monetize the cash-out fee.
Name the user: drivers who want same-day access to earnings instead of the weekly payout.
Build a metric tree: adoption, percent of eligible drivers who enable it; usage, cash-outs per driver per week; impact, hours driven and retention after adoption.
Add guardrails: fee revenue versus driver net earnings, and support tickets from failed transfers.
Segment by driver tenure and market, since part-time or new drivers likely rely on it more than full-time drivers.

Ties the north star to supply, hours driven in the days after a cash-out, not just transfer count.
Separates a vanity metric, total transfers, from a causal one, retention lift versus a holdout group.
Flags cannibalization: Instant Pay may just move up payouts drivers would have taken anyway, so tests transfer frequency against a control.

Treating transfer count alone as success without linking it to driver supply or retention.
Ignoring the cost side, card network fees against fee revenue collected.

How would you tell if Instant Pay is cannibalizing the weekly payout rather than adding value?
What would you do if fee revenue is strong but driver retention doesn't move?