How to Create a User Persona

There is a fast way to make a persona and a right way, and they share almost no steps.

The fast way: gather a few smart people in a room, invent a 34-year-old marketing manager named Sarah who does yoga and listens to podcasts, add a stock photo, ship it. It feels like progress. It takes an hour. And your team can smell it — which is why that persona ends up three folders deep in a shared drive, never opened again.

That is not a persona. It is a guess with a name tag.

The right way costs more — a week or two, and a handful of slightly uncomfortable conversations — and it produces something your team will actually reach for. This guide is the right way, broken into seven steps you can start today. If you want the theory behind it first, the complete guide to user personas covers what a persona is and why it works. Otherwise, let’s build one.

Step 1 · Frame the Decisions First

Don’t start with research. Start with the decisions the research has to serve.

Most guides tell you to jump straight into interviews. Skip that. Collect data before you know what it’s for and you end up with a pile of interesting facts and no idea what to do with them.

Ask one question: what will this persona help us decide? Redesign onboarding? Launch a new line? Stop sales from pitching every prospect the same way? Each answer sends you toward different people and different questions. A persona built to fix a product experience looks nothing like one built to sharpen marketing — and a persona built to do both usually does neither. If the person who pays isn’t the person who uses, decide that now; the split between buyer and user personas changes who you go talk to.

Write down three to five real questions the persona must answer. Not “who is our user?” Sharper:

• Why do some trial users convert in two days while others vanish?
• What’s the moment a new customer decides we’re worth paying for?
• What makes someone pick a competitor over us?

Those questions are your compass. Everything that follows serves them.

Step 2 · Gather Real Evidence

Watch what people do. It rarely matches what they say they do.

This is where most persona projects quietly die. Not because teams don’t know they should do research — because research feels slow and intimidating, so they skip it and invent instead.

You don’t need a six-month ethnographic study. You need conversations. Honest, slightly awkward conversations with people who use your product, tried it and left, or have never heard of you but match the profile you’re chasing. Talk to the ones who left. They tell you the truth the happy ones are too polite to mention.

How many? The Nielsen Norman Group recommends interviewing in rolling samples of about five users per group, continuing until each new conversation stops teaching you anything new. In practice that’s five to fifteen people. After the fifth, you’ll start finishing their sentences. That’s the signal you have enough.

Surveys, analytics, support tickets, and sales calls all help — but none replace the interview. A survey tells you what someone picked from a list you wrote. An interview tells you what keeps them up at night, in words you’d never have guessed. Here’s how the sources stack up:

Ask open questions that pull stories, not ratings. Not “do you like our product?” but “walk me through the last time you tried to solve this.” A few that consistently open people up: “What’s the most frustrating part of how you handle this today?” “Who else is involved when you decide?” “What would have to be true for you to switch?” Then use your best tool: silence. Ask, and shut up. The real answer lives in the pause. Our persona interview questions give you the full script, and the research methods guide covers recruiting, remote sessions, and how to triangulate what you hear against what your data shows.

Step 3 · Code the Notes, Find the Patterns

You’re not looking for the average customer. There is no average customer.

Now you have pages of notes, maybe transcripts. The instinct is to average it all out: “most users are 28 to 45, work in marketing, log in weekly.” That’s a demographic summary. It predicts nothing and rallies no one.

Do the opposite. Tag every quote and observation, group the ones that echo each other, and name the clusters that emerge. Researchers call it coding, then affinity mapping — and it’s the same move whether your notes live on sticky notes, index cards, or spreadsheet rows. You’re hunting for people who approach the problem in fundamentally the same way.

Signals you’ve found a real pattern: several people describe the same frustration in different words, a behavior shows up unprompted across interviews, or the same homemade workaround keeps reappearing. That workaround is gold — it’s a customer telling you exactly what your product should do next.

Demographics describe people. Behavior defines personas. When you can’t tell the difference, you build a census entry and call it a customer.

Step 4 · Cluster by Behavior, Not Demographics

Two people who look nothing alike on paper can be one persona.

This is the move that separates a real persona from a dating profile. Group people by what they’re trying to do and how they do it — not by age, income, or job title. Demographic similarity does not predict behavioral similarity. A 25-year-old freelancer and a 52-year-old enterprise director can share a frustration, a buying trigger, and a decision process. Age didn’t matter. The pattern did.

Most products settle into two to four distinct personas. Seven means you’re slicing too thin. One means you haven’t looked hard enough. The Nielsen Norman Group sorts the result into three types by rigor — lightweight, qualitative, statistical — and recommends the qualitative middle for most teams: real evidence, no research department required. If you’re still working from assumptions, that’s a proto-persona — useful, as long as you label it a guess and go validate it.

Step 5 · Draft the One-Page Persona

If a field doesn’t change a decision, it doesn’t belong on the persona.

Now you build the thing. A persona is a portrait, not a form — but it needs just enough structure to be scannable in thirty seconds. Here’s what survives the cut:

• A name and a face. “Would David understand this?” beats “would our target segment understand this?” every time. A name turns a spreadsheet row into someone your team can argue about.
• Goals. What is this person trying to accomplish — in their work or life, not just your product? Goals are the engine. Everything else connects back to them.
• Frustrations. What’s in the way, what’s broken about their current fix. These are your opportunities, wearing complaints as a disguise.
• A quote. One real sentence from a real interview. People argue with a summary. They remember a voice.
• A scenario. Three sentences of this person meeting your product in the wild — where they are, what they’re juggling, what happens next.

Resist the urge to over-design it. A gorgeous poster becomes too precious to change, and you will need to change it. Grab our free persona template for exactly this structure, and see finished ones in our persona examples across SaaS, e-commerce, and B2B. The first draft can be written for you in a couple of minutes — then you sharpen it with your own quotes.

Step 6 · Pressure-Test the Draft

Your first draft is wrong. That’s not a bug — it’s the plan.

A persona is a hypothesis. An informed one, built on real evidence, but a hypothesis all the same. So test it.

Start with your team. Show the draft to designers, engineers, salespeople, support agents — anyone who touches customers. Ask a plain question: “Does this ring true?” Your support team especially has a year of ground truth that never made it into a research doc. When someone says “this feels off,” dig in. Sometimes it’s a gap in your research. Sometimes you’ve stapled two personas into one, and you need to split them.

Then take the step almost nobody takes: show it to real customers who match the profile. “We wrote this description of someone like you — how close is it?” Their reaction sharpens everything, fast.

Step 7 · Put It to Work — and Keep It Alive

A persona in a slide deck is furniture. A persona in the workflow is a tool.

Here’s where most persona projects go to die. The research was solid, the document is handsome, and it goes straight into a folder after one all-hands. When the Nielsen Norman Group studied why personas fail, the autopsy barely mentioned the personas. The causes were organizational: created but never used, no leadership buy-in, no owner, left to go stale. Fix those and the method has never been the problem.

Making a persona live is unglamorous and simple:

• Name it out loud. “Would David find this?” in design reviews. “Which persona filed this ticket?” in support. A persona spoken weekly cannot die.
• Prioritize against it. Every feature proposal starts with “which persona is this for?” No answer is a red flag, not a detail.
• Give it one owner. Not a committee — one person who folds in new interviews and prunes what stopped being true.
• Review on a cadence. The NN/g treats personas as living documents, revisited every quarter or after each research round. Your customers keep changing; a persona that doesn’t is a photo of last year.

Do this and one set of personas starts working across the whole company — product prioritizes with them, marketing sharpens its message with them, sales recognizes who’s on the call. That’s what alignment actually looks like: not a document everyone signed off on, a person everyone can name.

The Whole Guide, in One Breath

Know what you need to decide. Talk to the people your decisions affect — especially the ones who left. Find the patterns. Cluster by what people do, not who they are on paper. Turn each cluster into one named, shareable page. Test it. Then use it every single day.

That’s how you create a user persona. Not by imagining your user — by meeting them. The effort is real: honest conversations, honest synthesis, the willingness to be wrong about who you thought your customer was. The payoff is bigger. A team aligned around one real person builds better products, writes better marketing, and wastes less time arguing about hypotheticals. The persona becomes the tiebreaker — the thing that turns “I think” into “we know.”

Frequently Asked Questions