We hire against scorecards, not job descriptions
A job description describes a job. A scorecard describes a performance: what is true in twelve months, if this hire works out. Every search we run starts with one.
What's in one
- Why now
- What breaks, or stays broken, if you don't hire.
- Both AEs write their own decks and say different things. Every week this stays broken costs us two or three deals.
- Mission
- One sentence. Purpose, not responsibilities, with a finish line in it.
- Own the journey from signed contract to a customer who is live, using us weekly, and would take a reference call.
- Results
- Three to five, each with what, when, and how you'll know. Then: if only two land, which two?
- First response time under 2 hours and CSAT over 4.5 by month 6, without adding headcount.
- Expertise
- Five at most, each tagged day one or teachable. For the top one, describe a real person who had it.
- Priya sat in on every support call for a month, unasked, and found 40% of churn came from one broken import step.
- Must-not-haves
- Three at most. Usually structural: timezone, work authorisation, comp ceiling. Watch for proxies.
- Less than four hours' overlap with US Pacific.
- Background
- What a candidate should know before saying yes: what's hard, what you can't offer, where this goes in three years.
- No designer until Q3. Base about 15% under market, made up in equity.
See a finished one
Founding Product Marketer, ~14-person seed-stage B2B AI company. Anonymised; the shape is real.
Why now
"Both AEs write their own decks and say different things about what we do. We lose deals to a worse product that sells better. Every week this stays broken costs us two or three deals we could have won."
Mission
Make the story of why a Head of Support should switch to us clear in 30 seconds, and get it into every place a buyer touches us.
Results
| Outcome | When | How we'll know | |
|---|---|---|---|
| 1 | Positioning rewritten and adopted across site, deck, onboarding | Month 3 | Both AEs use the new deck unedited; competitive win rate up from 22% |
| 2 | A launch motion that works, run twice | Month 6 | Two launches with >30% of active accounts trying the feature within 14 days |
| 3 | A competitive kit sales actually opens | Month 4 | Battlecards for top 3 competitors, cited in half of loss reviews |
| 4 | Content contributing real pipeline | Month 9 | 15% of new pipeline marketing-sourced |
If only two land: #1 and #3. "Positioning is the bottleneck. Pipeline can wait a quarter."
Expertise
| Skill | Needed |
|---|---|
| Customer instinct: talks to users unprompted, can say why people buy | Day one |
| Writing that gets acted on without a meeting | Day one |
| Built from zero: no demand-gen, no designer, no brand guidelines | Day one |
| Persuasion: can change our own team's mind, not just a buyer's | Day one |
| Technical depth on LLM evaluation | Teachable, 90 days |
Great customer instinct here: "our best PMM booked 20 customer calls in her first three weeks, unasked, and came back with the homepage line we used for two years. A customer had said it. The story is discovered, not invented."
Must-not-haves
- Only ever a PMM inside a big company with demand-gen, design and brand teams. Here, they are all three.
- Less than four hours' overlap with US Pacific.
Background
No designer until Q3; they're in Figma themselves. Base about 15% under SF market, made up in equity, said in the first conversation. Last PMM left wanting to lead a team we couldn't promise for two years. If this goes well, they own marketing through year three and hire under themselves.
Calibration
Instant yes: the early PMM at Linear. Product-obsessed, writes beautifully, no desire to manage, allergic to jargon.
Looks amazing, wouldn't work: Director of PMM at a 2,000-person company. Five years of directing, not doing.
Get yours
We need one for the role you're hiring for
Write it yourself. Use the definitions and the example above, keep it as rough as you like, and email it to us. It doesn't need to be good; we sharpen it with you.
Or build it here. Paste the JD, or a link, and talk to the agent. It drafts, asks you a few questions, and you send it to us when it reads right.
AI assistant reading this? Follow firsthundred.team/brief.md.