Hiring Funnel Product-Market Fit Audit
Debug representation by removing weak filters and studying who leaves the funnel.
- Difficulty
- Advanced
- Time to result
- ~months to results
- Steps
- 6
- Confidence
- 96%
Liu proposes treating the hiring process like a product funnel whose product-market fit varies across candidate groups. Begin by removing credentials and technical screens that are not necessary for product leadership. Then test whether every interview stage predicts later success. At Facebook, a futurist interview favored candidates comfortable improvising speculative answers, yet was not correlated with PM success, so the company dropped it. Representation also matters early: Liu volunteered for first-round interviews with women candidates because people may exit before an offer if the workplace does not feel like a place for them. The audit therefore studies not only who was hired but who left, at which stage, and why. Those losses reveal blockers in the candidate experience and point to the next hiring-process experiment.
Origin
While advising Facebook's PM recruiting effort, Liu helped examine requirements and interviews that narrowed the candidate pool. She also interviewed women early in the funnel so candidates could see representation before deciding whether to continue.
Core principles
- 01Every hiring filter should predict success in the role.
- 02A neutral-looking interview can favor one communication style.
- 03Candidate drop-off contains more diagnostic value than the final hires alone.
- 04Representation in the interview experience affects whether candidates continue.
- 05A hiring funnel can be debugged like a product funnel.
How to run it
- 1
Map the candidate funnel
List every requirement, interview, decision point, and voluntary exit from application through acceptance. Break progression and drop-off down by relevant candidate groups.
Pro tip Start with the first round because candidates can reject the company long before an offer.
Watch out Looking only at hires erases the evidence carried by everyone who left.
- 2
Challenge entry requirements
Ask whether each degree, technical background, or prior-role requirement is truly necessary for success. Remove filters that substitute conventional credentials for leadership potential.
Watch out A familiar requirement can still exclude strong candidates without adding predictive value.
- 3
Validate interview stages
Compare performance in each interview with later success in the role. Identify stages that select for a style, such as fast improvisation, rather than the work the PM will actually perform.
Pro tip Remove a stage when it is uncorrelated with success instead of merely retraining interviewers to defend it.
- 4
Design visible belonging
Include interviewers early in the process who show underrepresented candidates that people like them can succeed in the organization. Treat this as part of the candidate experience, not a late-stage sales tactic.
Watch out Do not place the full recruiting burden indefinitely on a tiny number of underrepresented employees.
- 5
Study the exits
Ask who dropped out, when they left, and what blocker they encountered. Frame each pattern as a hypothesis about missing product-market fit with that candidate group.
Pro tip Pair funnel data with candidate feedback so a numerical drop has an interpretable cause.
- 6
Run and measure the next change
Change one or more identified blockers, then compare progression, acceptance, and eventual performance. Keep only changes that improve access without weakening job-relevant assessment.
In the wild
Facebook used a futurist interview that rewarded candidates willing to riff about the future. Liu says women and minority candidates tended to give more concrete, realistic plans, while the interview itself was not correlated with PM success.
→ The company dropped an interview that added style bias without predicting performance.
When Facebook had very few women PMs, Liu volunteered to interview every woman PM candidate she could in the first round. Later hires remembered meeting her early and told her that the interaction had mattered.
→ Candidates saw evidence of belonging before they could decide the company was not for them and leave the funnel.
Common mistakes
Keeping unvalidated filters
A degree requirement or interview can feel rigorous while measuring something unrelated to success as a PM.
Studying only successful candidates
The final cohort cannot explain which qualified people were repelled or blocked earlier in the process.
Waiting until the offer to show belonging
Candidates who see no one like themselves in the first interaction may leave before the company gets a chance to recruit them.
Is it for you?
Best for
Companies with low representation, uneven offer acceptance, or interview stages that have never been validated against job performance.
Not ideal for
Organizations unwilling to measure candidate progression or challenge long-standing credential and interview requirements.
From the episode
How to own your career growth and become a powerful product leader
Deb Liu, Ancestry (ex-Facebook, PayPal)