Hire

Turn a fuzzy 'we need someone' into a role scorecard, a loop where every interview has a job, a reference guide and a close plan. Built from 586 insights from 210 Lenny's Podcast guests. Lead team: Adam Ward (Cursor, ex-Pinterest/Facebook recruiting), Keith Rabois (Khosla, PayPal/Square), Lauren Ipsen (executive recruiter, ~85 senior placements).

When to use

• You are opening a role (PM, designer, engineer, marketer, first sales hire, head of product) and have no written definition of great.
• You have a job description or interview loop and suspect it tests interview skill, not job skill.
• You are about to run reference checks and need questions that get past polite praise.
• You are hiring the first person in a function (first PM, first marketer, first growth hire, first AE) and don't know the archetype.
• A strong candidate is wavering, or you keep losing finalists, and you need a close plan.
• You made a hire 30 days ago and want to know if it was right.
• You are hiring in the AI era and unsure about AI-allowed interviews, juniors, or generalists.

Step 1 — Diagnose (ask before answering)

If the user attached a JD, scorecard, loop doc, or candidate notes, read them first and only ask what is still missing. Max 4 questions:

1. What problem must this person solve in the next 12 months, and for what stage/team? Routes to the scorecard (Lowenhar's 12-month test, Ipsen's best-for-this-role-at-this-time). If they can't answer, the first deliverable is the outcome, not a posting.
2. First in the function, or joining an existing team? First hires route to archetype plays (pi-shaped marketer, pioneer sellers, growth competency model); existing teams route to team-gap hiring.
3. How many strong candidates can you realistically reach, and how much founder/manager time is there? Routes between the top-50 executive-search model (low volume, high stakes) and a structured funnel (high volume).
4. Can the job be simulated in 2 hours to 2 days? Routes to work samples or paid trials versus a reference-heavy loop (senior/exec roles).

Step 2 — Pick the play

Step 3 — Run it

A. Write the scorecard (always do this first)

1. Outcome statement. Write: It is 12 months from now and this hire has been great. What changed in the business? (Lowenhar). Name 2-3 outcomes with numbers. Add the mandate and what 'crushing it' looks like at 90 days, 1 year, 2 years (Ipsen).
2. Why this role, why now. If you cannot name the specific problem, delay the search (Ipsen: board-driven searches fail from inception). If the CEO still wants to own the roadmap, don't hire a CPO (Manik Gupta).
3. Archetype and spikes. Decide where the hire must major, minor and spike. Use a taxonomy that fits: product leaders as platform / core-consumer / growth-monetization (Ipsen); PMs as consumer, growth, GM, platform, research/AI (Peter Deng); execs as architect, optimizer, scaler (Lowenhar: hire an architect only if they've been one). Individuals spike; the team is well-rounded (Mihika Kapoor, Eeke de Milliano's six-month strengths/weaknesses chart).
4. Four to five ranked skills plus attributes and values (Oji Udezue via Bridgewater: skills, attributes, values). Stack-rank them; the top two get a dedicated interview each (Ward).
5. Level test: P, S, H, E. Is this person handed the Problem, Solution, How, or do they define the problems? (Shishir Mehrotra's PSHE). Write which one the role needs.
6. Anti-requirements. Replace 'five years of experience' with the concrete capability it stands for (Bob Moesta). Strike company-logo requirements; define great by transferable skills (Ward). For a startup, check when the candidate joined their big-name employer and at what size (Meltem Kuran Berkowitz).
7. Write the JD as an experience: context, what the role is about, what progress means, how you reward it (Moesta). Open with the team mission (Ebi Atawodi). State intensity honestly so people self-select (Anton Osika, Patrick Campbell).

B. Source (pick by volume)

Low volume / high stakes (default for senior, first-in-function):

1. Build the top-50 list. Ask trusted high-taste people about a specific trait, not 'best engineer you know'; add names mentioned by two or more (Ward).
2. Market-map the best companies for the function, then go a layer deeper to who was breakout versus riding the wave (Ipsen). To hire a function leader, find 3-4 best-in-class companies for a similar customer and recruit the lieutenant, not the head (Rajaram).
3. Keep the next conversation going: for 'not now', ask for coffee or a teammate intro, not an interview. Warm intro beats cold message (Ward). Court over months with no agenda, as advisers (Ipsen: a seven-month courtship).
4. Run network-mining sessions with employees every other week for about six weeks, with follow-up accountability (Ward). Founders: budget about two hours a day (Rajaram); up to 30-40% of time for early teams (Jason Shah).

High volume: structured loop (Step C); screen with a written job-simulation doc (Kazanjy: 57% don't finish it); keep 'founder, even failed' explicitly welcome (Ayo Omojola).

C. Build the loop: one signal per interview

1. Hiring manager first conversation is the highest-leverage step; allow 3-4 conversations for a hard-to-land candidate (Ward). Hiring manager decides; founders stamp; debrief same or next day.
2. Work sample or paid trial where feasible: candidate picks 1 of 3 problems, about 3 hours (Merci Grace); two-day onsite mini-project (Michael Truell); paid week for finalists (Jason Fried); paid multi-day trial (Karri Saarinen, Elena Verna). Costly, so keep for finalists.
3. Candidate presentation: half free choice, half your prompt; or the target company's real problem with a prep call where you give feedback and watch whether they use it (Mehrotra, Melissa Tan).
4. Question mix: home, away, neutral court (Mehrotra). Mostly neutral-court (made-up problems), the presentation as the away-court slot, almost no home-court strategy questions.
5. Rubric before intuition (Annie Duke, Kahneman): write the abstract rubric, build structured questions from it, then apply gut. Experienced interviewers may trust intuition if they decode it into named traits afterward (Matt MacInnis' SPOTAC).
6. Question bank (pick one per competency, ask the same set to every candidate):

7. Never test: structure-sounding answers whose ideas are weak (Jackie Bavaro), rehearsed behavioral stories (Nikhyl Singhal's 'what is conventional wisdom you think is hogwash' breaks rehearsal), or magic-wand solutions in PM work samples (Merci Grace).
8. AI-era loop: allow AI tools and see what they build in an hour (Brendan Foody), then bring them onsite to watch them reason unaided (Varun Mohan). For product roles screen curiosity and agency over prior AI experience (Nick Turley, Logan Kilpatrick, Kevin Weil).

D. Run references (the most valuable guide you write, per Mehrotra)

1. Run them early so no one is dinged after a full loop (Mehrotra). Provided references nearly always come back positive (Ipsen: two negatives in a career), so weight backchannel references: peers and direct reports beyond the list. Silence on achievements is a negative signal.
2. Do many for senior hires and keep going until you hit a negative one; frame the question around the capability you need, not 'was he a good employee' (Rabois: Tony at DoorDash does about 20).
3. Ask: Would you hire them again? Would you report to them? Would you put your name on an endorsement? (Ipsen); Would you enthusiastically rehire for that role? 1-10? Top 1% or top 10%? (Halligan); What feedback will I be giving them in their first performance review? (Paul Adams).
4. Blind version (Mehrotra): don't name the candidate or reveal your ranking. On this team, who regularly identifies the right problems? Who is most reliable on the hardest ones? Offer four neutral personas (problem finder, creative solver, mobilizer, playbook executor).
5. Cut-through trick: tell the referee you're about to hire and ask them to call back only if not hiring would be a huge mistake (Phyl Terry). No callback is data.
6. Treat each datapoint with a grain of salt, write it down, weigh holistically. For mixed 'decent' reviews, find people who worked closest (Ipsen).

E. Decide and close

1. Debrief: same or next day; hiring manager decides, with input, not a committee (Ward). If panels average toward the candidate with the fewest weaknesses, hire the spiky one instead (Halligan). In hypergrowth add a Bar Raiser who runs the debrief against criteria (Bill Carr; budget up to 10 hours a week).
2. Move fast once ready: hiring manager conversation within an hour of interest; same-day scheduling (Ward: slow to start, fast to close).
3. Ask how they will choose: You'll have several strong offers; how specifically will you decide? Pitch on their answer (Will Larson). Pitch your blocker to their skill so they bet on themselves (Rabois).
4. Team sport: daily 10-minute standup and a Slack channel per top candidate; every exec takes a call; involve spouse/family; a low-cost personal gift at the offer meeting (Ward, Luc Levesque).
5. Unsell option (Kevin Yien): at offer, send an email of at most eight bullets listing the hard truths and fears you heard. Expect to lose some (he lost about 30% at first); the ones who stay are A-plus. Never promise titles, stability or 'I'll never hire above you' (Molly Graham).
6. Offer-to-start: dinners, early laptop, weekly touchpoints (Ward).

F. Close the loop

Set a calendar entry 30 days out: knowing what I know now, would I make the same decision? Ask the whole hiring team (Rabois, Uri Levine). If no, act. Then review at 6/12/18 months against the level and scorecard to find holes in the loop (Yien). Nobody is right more than about half the time (Molly Graham), so fire as skillfully as you hire.

Where the experts disagree

1. Do interviews predict performance? Camp A: structured loops with rubrics (Annie Duke, David Singleton, Bill Carr, Claire Hughes Johnson) vs Camp B: do the work instead (Farhan Thawar, Karri Saarinen, Michael Truell, Nir Eyal, Kristen Berman, Adam Ward). Use A when volume is high, the work can't be sampled, or the role is senior and the best proxy is references. Use B for IC roles with observable output and under 20 finalists. Default: rubric loop plus a paid or timed work sample for finalists, with references outranking both for senior roles (Mehrotra's order: references, then work product, then interviews).

2. Experience or hunger? Camp A: hire people who've done it (Ben Horowitz, Melissa Perri's mix of grown and hired, Pete Kazanjy for sales) vs Camp B: promote from within and hire raw talent (Keith Rabois, Gokul Rajaram for the first PM, Brian Halligan, Dmitry Zlokazov, Paul Adams, Laura Modi). Use A where trial-and-error knowledge is codified (global sales org, security, a function you can't coach). Use B for value creation, ambiguity, or when you can't outbid. Rabois: hire outsiders for value preservation, not value creation. Default: homegrown plus a few proven outsiders (Halligan's Red Sox mix), and never a big-company hire without asking when they joined.

3. Sell the candidate or talk them out of it? Camp A: sell and out-care (Adam Ward's caring is free, Levesque, Scott Wu) vs Camp B: unsell (Kevin Yien, Josh Miller, Patrick Campbell). Use A on the work, the mission and the person's growth. Use B on chaos, pace and culture. Do both: sell hard on the problem, state the downsides before the offer, because mismatches that leave in six months are the expensive failure.

4. Hire slow or fast? Camp A: slow and bar-obsessed (Grant Lee, Brendan Foody for the first 10, Halligan) vs Camp B: you hired too slowly (Michael Truell, Foody after PMF). Slow before product-market fit and for the first 10; fast once demand exceeds capacity. Don't set headcount targets (Grant Lee).

Deliverable

# Hiring Kit: <Role>, <Team>, <Date>

## 1. Scorecard
- Why this role, why now (one sentence):
- 12-month outcomes (2-3, with numbers):
- Mandate; 90 days / 1 year / 2 years:
- Archetype, major / minor / spike:
- Level (P / S / H / E):
- Ranked skills (4-5) | Attributes | Values:
- Not required (replaces 'N years', logos):
- Red flags for this role:

## 2. Sourcing plan
- Top-50 list owner, source of trait-specific referrals, 3-4 best-in-class companies:
- Warm-intro path for top 10; weekly cadence for 'not now':

## 3. Loop
| Stage | Owner | Signal tested | Exercise or question | Pass bar |
|---|---|---|---|---|
| HM conversation | | | | |
| Work sample / trial | | | | |
| Presentation (prep call?) | | | | |
| Competency interviews (1 per ranked skill) | | | | |
| Culture / values (who qualifies?) | | | | |

## 4. Reference guide
- Who (3+ backchannel), when (before onsite or before offer):
- Blind questions; hard questions; success-case framing:

## 5. Decision rule
- Who decides; debrief within 24h; veto or Bar Raiser; spiky or fewest-weaknesses:

## 6. Close plan
- Candidate's top 3 decision factors; who calls; personal touch; unsell email:

## 7. After the hire
- 30-day rehire check (date); 6/12/18-month review (dates):

## Next 3 actions
1.
2.
3.

Grade existing work

Score a JD, scorecard or loop on 1 / 3 / 5, then total out of 40 and list the top 3 fixes.

Fixes come from the lowest scores first; cite the guest and framework behind each.

Red flags

• 'I'll know it when I see it' (Ward). You will hire your own reflection.
• The funnel of doom: reach out to 100, filter down to the people you caught on a good day (Ward).
• Board-mandated search with no named problem (Ipsen); the unicorn head of product (Ipsen).
• Hiring for logos or years (Moesta, Ward, Austin Hay on false resume shortcuts, Lowenhar).
• Hiring a VP of Sales before two reps hit quota (Lemkin); hiring a perfect-on-paper big-company seller who consumes your scarce leads (Lemkin).
• Hiring the big-company exec into a 50-500 person startup (Halligan, Dmitry Zlokazov, Paul Adams).
• Interviewing for sameness (Seth Godin); polished structure with weak ideas (Jackie Bavaro).
• Recruiter incentives set the bar (Yien): pass-through and offer-close rates push doubtful candidates through. Recruiters are a confidence engine, not a decision engine (Ward).
• Skipping or delaying references (Singleton, Mehrotra), and trusting only the list they gave you.
• Hiring fast to hit a headcount number (Ray Cao, Grant Lee).
• Promising what you cannot control (Molly Graham).
• Waiting for the lagging indicator on a new seller: nine months in you learn what two months of leading indicators would have said (Kazanjy).

Receipts

"By definition, that is not the top 20%. That is just the 20 people who you caught on a bad day for them. Good day for you, bad day for them." — Adam Ward, Lenny's Podcast (00:11:59)

"Same person, wrong question, wrong result." — Keith Rabois, Lenny's Podcast (00:12:43)

"Just because this is the best talent, that doesn't necessarily mean they are the best talent for this role today." — Lauren Ipsen, Lenny's Podcast (00:12:46)

"Reference check at the top, work, product and presentation next, and then all the interviews. When they disagree, that's the order we judge." — Shishir Mehrotra, Lenny's Podcast (01:23:23)

"in six months, if I'm telling you what to do, I've hired the wrong person." — Peter Deng, Lenny's Podcast (01:02:03)

"Interviews are not a good predictor of performance. We know this." — Farhan Thawar, Lenny's Podcast (01:03:28)

Go deeper

• references/frameworks.md — every framework in this skill, how to run it
• references/quotes.md — verified quotes with timestamps
• Related skills:
  • exec-comms — when the hire is an exec and you need to write the offer narrative or role memo.
  • hard-conversations — when the 30-day check says no, or you must give feedback to a candidate or a struggling hire.
  • b2b-sales — for the first-sales-hire motion, ICP and pipeline beyond the hiring plan.
  • career — when you are the candidate, to run the same loop in reverse and evaluate the employer.