Sales Enablement Aside—Buyer Persona Development: How to Build B2B Personas That Actually Sharpen Messaging and Targeting
By Rick Elmore ·
Every few months a prospect sends me a persona deck they paid an agency to build. It's usually gorgeous. Stock photo of "Marketing Mary," a fake quote, her favorite apps, whether she prefers email or Slack. And it's completely useless for writing a single cold email or building a target list. Nobody on the sales team has opened it since the kickoff call.
That's the problem with how most B2B teams approach buyer persona development. They treat it as a branding exercise instead of a targeting and messaging exercise. A persona should change what you say and who you say it to. If it doesn't do both, it's decoration.
Here's how we build personas that actually feed the revenue engine.
- Personas and ICP are not the same thing. ICP tells you which companies to target; personas tell you which humans inside them to move and how.
- The only persona fields worth capturing are the ones that change a sentence or a filter. Role, pains, triggers, objections, and the language they use. Skip the hobbies.
- Research beats imagination. Win/loss calls, sales call recordings, and real prospect language beat a whiteboard brainstorm every time.
- AI accelerates persona research, but it can't invent your buyers. Feed it your real data; don't let it hallucinate a customer.
- A persona is only finished when it produces messaging and a targetable list. That's the test.
Persona vs ICP: stop conflating them
This trips up more teams than anything else, so I'll be blunt about it. Your Ideal Customer Profile describes the account: industry, company size, tech stack, revenue, growth stage, the triggers that make a company a fit. Your persona describes a person inside that account: their role, what they're measured on, what keeps them up at night, and what makes them say no.
You need both, and they operate at different layers of your outbound. ICP decides which 2,000 companies land on your target list. Personas decide who inside those companies you contact and what the message says. A CFO and a VP of Sales at the same fitting company get radically different emails, because they care about different outcomes and fear different things. One ICP, multiple personas.
When teams skip the distinction, they write one generic message "to the company" and it lands with nobody. Or they build detailed personas with no account criteria and their reps waste time on individuals at companies that were never going to buy. Get the layers straight first.
| Dimension | ICP (the account) | Persona (the person) |
|---|---|---|
| Answers | Which companies do we target? | Who inside them do we move, and how? |
| Example criteria | Industry, headcount, revenue, tech stack, funding, growth signals | Role, goals, pains, buying triggers, objections, vocabulary |
| Drives | List building and account selection | Messaging, sequencing, and content angles |
| Fails when | Too broad, reps chase bad-fit accounts | Too decorative, nobody uses it to write copy |
The only five things a B2B persona needs
Cut everything that doesn't change a message or a filter. I've never once written a better subject line because I knew the persona's favorite podcast. Here's what earns a spot.
Role and mandate. Not just the title, but what this person is actually accountable for and how they're measured. A "VP of Marketing" measured on pipeline sourced writes a different reality than one measured on brand awareness. Titles vary wildly across companies, so capture the function and the metric, then map the range of titles that carry it. That title list becomes a literal filter in your list-building tool.
Pains. The specific, recurring problems tied to their mandate. Be concrete. "Wants to grow" is useless. "Sales reps spend half their week on manual research and CRM updates instead of selling" is something I can write a whole sequence around. Good pains are usually operational and specific enough that the buyer nods before you finish the sentence.
Triggers. What makes the pain urgent now. A new funding round, a leadership hire, a missed quarter, a competitor's move, a hiring spike, a tool migration. Triggers are the bridge between persona and timing, and they're where outbound goes from cold to relevant. Most personas ignore them entirely, which is why so much outbound feels like it arrived at random.
Objections. The real reasons they don't buy or don't respond. "We already have a tool for this." "No budget until next quarter." "I don't trust AI to touch my pipeline." If you know the top three objections per persona, you can pre-handle them in your messaging and arm your reps for the call. This is the field that most directly separates a persona that helps close deals from one that just sits in a deck.
Language. The exact words this person uses for their problem. Buyers rarely describe their pain in your product's vocabulary. Capturing their phrasing, from real calls and posts, is what makes copy sound like it was written by someone who lives in their world. This one field is the difference between "leverage our omnichannel engagement platform" and "stop your reps from drowning in busywork."
Where the research actually comes from
You cannot brainstorm your way to a good persona. Everything above has to come from evidence, or you're just projecting your own assumptions onto a stock photo. Here's the hierarchy of sources I trust, roughly best to worst.
Start with your own sales calls. If you record calls, you're sitting on the richest persona research available. Every discovery call is a buyer telling you their pains, their triggers, and their objections in their own words. Read or listen to twenty of them from your best-fit deals and patterns emerge fast. Win/loss conversations are even better, because losses tell you the objections you're not handling.
Next, talk to your reps and CSMs directly. They carry pattern recognition that never makes it into a CRM field. Ask them: what does this buyer always push back on? What makes them light up? What's the phrase they keep using? Twenty minutes with a good AE beats an hour of desk research.
Then mine what buyers say in public. LinkedIn posts, community threads, podcast appearances, job descriptions for the role you're targeting. Job posts are underrated: they spell out exactly what a company thinks a role is responsible for and what problems they're hiring to solve.
Only after all that should you touch third-party data and industry reports, and treat them as context, not gospel. The goal is a persona grounded in what your actual buyers say and do, not in what a generic report claims the category cares about.
How to use AI without inventing a customer
AI is genuinely useful for persona development, but only if you point it at real inputs. The failure mode is asking a model to "create a buyer persona for a VP of Sales" and accepting whatever plausible-sounding fiction it returns. That's not research. That's a confident guess dressed up as data, and it will lead your messaging straight into the generic middle where everyone else lives.
The right way: feed the model your evidence and let it find patterns faster than you could by hand. Paste in transcripts from ten discovery calls and ask it to cluster the recurring pains and objections. Drop in fifteen job descriptions for a target role and ask what responsibilities and problems show up most. Give it a batch of your buyers' LinkedIn posts and ask it to pull out the exact phrases they use to describe their challenges. Now the AI is doing analysis on your real data instead of hallucinating a customer.
From there, AI helps you go from persona to output. Once you've got a tight, research-backed persona, it can draft messaging variants per persona, generate objection-handling snippets your reps can pull into calls, and produce content angles that match each buyer's pains and language. This is where personas stop being a document and start being infrastructure. We wire this directly into the outbound and content systems we build, so the persona's pains and language flow into the actual sequences reps send. That's the whole point of an integrated revenue engine rather than a stack of disconnected tools.
The test: does it produce messaging and a list?
Here's how I know a persona is done. Take it to two people. Hand it to whoever builds your target lists and ask: can you turn this into filters? The role, the title variants, the company criteria from your ICP, the triggers you can detect with signal data. If they can build a list from it, the targeting half works.
Then hand it to whoever writes your outbound and ask them to draft a sequence using only what's in the persona. If the pains, triggers, objections, and language give them everything they need to write something specific and human, the messaging half works. If they come back asking "but what does this person actually care about," the persona isn't finished.
A persona that passes both tests earns its place. One that only produces a nice slide gets deleted. I'd rather have a rough, ugly one-pager per persona that reps actually use than a polished deck nobody opens. If you want to see how we build this into a full lead-gen and sales system rather than a standalone deliverable, our packages lay out where persona work fits.
Build two or three personas to start. Not eight. Most B2B companies have a small number of people who actually drive and block deals, an economic buyer, a champion, maybe a technical evaluator. Nail those, get them into your messaging and targeting, and expand only when the data tells you to.
Frequently asked questions
How many buyer personas should a B2B company have?
Fewer than you think. Start with two or three: the economic buyer who controls budget, the champion who feels the pain daily, and sometimes a technical or operational evaluator. Building eight personas usually means you're describing job titles, not distinct buying behaviors. Add personas only when real deal patterns show a genuinely different set of pains, triggers, or objections.
How often should we update buyer personas?
Treat them as living documents, not annual projects. Revisit them every quarter using fresh call recordings and win/loss notes, and update immediately when something shifts: a new competitor, a market change, a product launch, or a pattern of objections your reps keep hitting. Personas drift out of date quietly, and stale ones quietly degrade your messaging.
Can AI build buyer personas for us?
AI can accelerate the work, but it can't invent your buyers. Point it at real inputs, call transcripts, job descriptions, buyer posts, and it will cluster pains, surface objections, and extract the language your buyers actually use faster than any human. Ask it to build a persona from nothing and you get a plausible fiction that will hollow out your messaging. Real data in, useful patterns out.
If your personas are sitting in a deck nobody uses, let's fix that by wiring research-backed personas into your outbound targeting and messaging. Book a Revenue Systems Audit.