Metrics question

How would you solve homelessness in downtown San Francisco?

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

Structured problem-solving on an open-ended civic issue, testing whether the candidate can scope a huge problem into a tractable, prioritized plan.

Clarify scope: answer as leading a specific funded initiative, since a PM interview expects a scoped plan, not policy invention.
Segment the homeless population by root cause, mental health or addiction, job loss, housing cost, chronic versus situational, since one intervention won't fit all.
Prioritize the most tractable segment first, like situationally homeless individuals needing housing-cost assistance and job placement, since impact there is faster and more measurable.
Propose a concrete, resourced intervention for that segment, like rapid rehousing plus job-placement partnerships, not a vague call for more shelters.
Define success as individuals moved into stable housing per quarter, and a guardrail against merely displacing the problem to another neighborhood.

Segments the population instead of treating homelessness as one monolithic problem, the key insight interviewers look for.
Picks the most tractable segment to prioritize, showing resource-constrained judgment rather than an unbounded wish list.
States a measurable outcome metric and a guardrail against displacement.

Proposing generic, unscoped solutions like 'more affordable housing' without segmenting or prioritizing.
Not acknowledging resource constraints, producing a plan the asking entity, a company not a government, can't actually execute.

How would you measure success without just displacing the problem to another neighborhood?
Which stakeholders would you need to align first?