Recruiting for Persona Interviews: Finding People Who Will Tell You the Truth
A persona is only ever as good as the twelve or fifteen conversations underneath it, and those conversations are decided by recruiting, the least glamorous and most consequential step in the whole exercise. Recruit loosely and the interviews fill with professional respondents who have been in six studies this month and have a polished opinion about everything and a real workflow about nothing.
Write the screener from the decisions, backwards
Start with what the persona must decide: pricing, onboarding, a feature fork, a segment boundary. Each decision names who holds the answer. If the fork is whether small teams adopt without IT support, the screener needs people who run small teams and made a software decision in the last year, not people who match a demographic sketch of a small team leader.
Screen for behaviour and recency, not for resemblance. Did you buy, implement, or reject a tool like ours in the last twelve months is a fact. Do you consider yourself a tech savvy decision maker is a performance. Recruiting on facts produces interviews full of specific stories. Recruiting on self description produces interviews full of adjectives, and personas built from adjectives persuade nobody who has met a customer.
Exclude the professionals politely
Every research panel contains people who do studies for income. They are fluent, agreeable, and catastrophic for persona work, because their skill is sensing the answer the interviewer wants. Screen them out with specifics a genuine user answers instantly and a professional fakes badly: the exact plan tier they are on, the last task they did in the product, the workaround they use for the thing that annoys them.
Use your own customer list first, panels second, and snowball referrals, ask each good interviewee who else does this job your way, as a targeted supplement. The best persona participants are usually one introduction away from someone you already interviewed, because real practitioners cluster.
Numbers, mix, and the saturation test
Eight to twelve interviews per segment is the working range, and the signal to stop is saturation: new interviews stop producing new jobs, new objections, and new language. If interview ten sounds like interview three with a different name, the segment is mapped. If it keeps surprising you, the segment is probably two segments wearing one label, which is a finding, not a failure.
Protect the mix inside the segment. Deliberately include the newest customer and the longest tenured, the enthusiast and the reluctant user dragged in by their employer, because a persona averaged from fans describes a customer who does not need convincing, in a market where convincing is the job.
Pay properly and say what it is for
Incentives should respect the hour and the seniority you are asking for. Underpaying filters your sample toward people whose time is cheap, which quietly biases the persona toward a customer your pricing will never target. Tell participants plainly that the conversation will shape how the company describes and builds for people like them. People who know what the interview is for give straighter answers than people guessing at a hidden agenda.
Record, transcribe, and tag the screener answers to the transcript. Six months later, when someone challenges the persona in a meeting, the recruiting trail is the defence: who was asked, how they were chosen, what they actually did. A persona that can show its recruiting can survive scrutiny. One that cannot is a poster.
Keep reading
Related guides on this site: Handing Personas to Sales and Support: Making Research Usable on Calls Micro Personas for Niche Audiences: When a Broad Persona Hides the Buyer Practical answers about personalization.
Frequently asked questions
How many interviews does a persona need?
Eight to twelve per segment is typical, stopping when new interviews stop adding new jobs, objections, or language. Continued surprises usually mean the segment contains two different customers and should be split.
What is the most important screener question?
One that verifies recent, real behaviour: a purchase, implementation, or rejection decision in the category within the last year. Behaviour and recency predict a useful interview far better than demographic resemblance or self described expertise.
Why exclude frequent research participants?
Professional respondents are skilled at sensing and supplying the expected answer, which poisons persona work that depends on unpolished, specific accounts of real workflows. Screen for product specific facts only a genuine user would know.
Should persona participants be paid?
Yes, at a rate that respects their time and seniority. Underpaying biases the sample toward people unlike the customers the business actually sells to, and the persona inherits that bias invisibly.