Build a Buyer Persona From First-Party Data
It’s Tuesday, 4:30 p.m. You need a real buyer persona by Friday—grounded in conversations, not adjectives. You’ve got call recordings, a few win-loss notes, and a founder who’s seen every deal. Here’s how to turn that pile into a living persona—and keep it current without babysitting it.
- Gold standard: dedicated research interviews
- Good: Gong or win-loss calls (bias exists, insights still strong)
- In a pinch: a 20-minute interview with your best SME or founder
The moment you actually need a persona
A launch is coming. Sales wants talk tracks. Creative wants buyer language. You open the old PDF and realize it describes last year’s market. The fix isn’t a new template; it’s first-party data—what people said in interviews and calls this month.
Start with first-party data
Use what you already have. Interviews and call recordings surface the details that move decisions: exact phrasing, triggers, and objections. If interviews aren’t possible this week, interview a strong internal SME or the founder. The point is to ground the persona in real words, not guesses.
Crawl – ChatGPT (ship by Friday)
What you do
Paste 5–10 interview or call summaries and a quick SME/founder interview into ChatGPT. Ask for pains, goals, triggers, objections, and verbatim buyer language. You’ll ship a one-page persona today; updates are manual.
Copy-paste prompt (use as-is)
It's an 11-page prompt.. so we put it in a Google Doc here to make this page more readable.
Walk – n8n / Gumloop (keep it fresh weekly)
What you do
Automate a light weekly refresh so the persona stays current without copy-paste.
Flow outline
- Fetch new call transcripts and win-loss notes
- Summarize to fields: pains, triggers, objections, buyer words
- Light external checks: LinkedIn role pages, competitor “Why Us” pages
- Write back: update a Google Sheet and, if comfortable, CRM fields
- Notify: post a weekly summary in Slack/email
Output
Updated persona fields and reusable snippets every week. Caveat: no-code flows are rigid—when fields or pages change, expect to adjust.
Run – Operating agents (reasoning + ops)
Agents research, reason, and operate inside your tools. They reconcile title variants (“CISO” vs “Chief Information Security Officer”), adapt when layouts move, update CRM/MAP persona fields, and halt on uncertainty. Every change is logged.
Output: a living persona synced across systems with QA and an audit trail—no stitching.
Fields that matter
- Top pains — concrete problems buyers name
- Goals — outcomes they’re measured on now
- Decision triggers — audit, migration, missed target, new boss
- Objections — cost, bandwidth, risk, “we already have X”
- Buyer language — 5 exact quotes worth reusing
- Related roles + title variants — who else says yes/no
One-page outline
- Who — title family, team size, reports to
- Why now — the trigger that makes action reasonable
- Top pains (3) — in the buyer’s words
- Goals (2–3) — outcomes that matter this quarter
- What they need to see — proof/risk reducers
- Message angles (3) — one-liners to test next week
- Buyer language (5 quotes) — verbatim
- Title variants — your dictionary for clean targeting
Title nuance: CISO vs “Chief Information Security Officer”
CRMs hide duplicates and near-matches. Your list builder might miss “Head of Security” when you filter for “CISO.” Keep a title dictionary (canonical role → acceptable variants) and apply it anywhere you build lists or logic. Agents at Run can maintain and apply it automatically.
Get started
- Gather 5–10 call or interview summaries
- Paste into ChatGPT with the Crawl prompt and ship the one-pager
- Stand up the Walk workflow so fields refresh weekly
- Add Run when you want agents to research, reconcile titles, update fields across tools, and keep an audit trail
Ship something real by Friday. Improve it every week.
