SEO tools buyer personas: data sources that earn their place

What to take away

Evidence worth collecting

Most persona templates ask for job title, age and budget. Those fields predict very little. The inputs that change a brief describe a search behaviour, a trigger and a constraint. A trigger might be a new head of search joining. A constraint might be a licence that renews in March.

Group the evidence into four types: demand, survey, operational and competitive. Each type answers a different question. Mixing them without labels produces personas that read well and forecast badly.

Demand data in practice

Keyword exports show what people search before they buy. Pull terms around "rank tracker", "SERP monitoring" and "SEO reporting" into a spreadsheet. Tag each term by intent: comparing suppliers, checking a feature, or describing a problem with no product in mind. Those tags become your segment boundaries.

For keyword ideas and historical metrics, use Google Keyword Planner. You can also pull matching terms and SERP information from Ahrefs Keywords Explorer, keyword and intent data from Semrush Keyword Magic Tool, or suggestions from Moz Keyword Explorer. Record the term, country, metric and intent where available, keeping each tool's figures separate because their measures may differ.

Use Google Search Console's Performance report to check your own site's organic demand. Export queries, clicks, impressions, CTR and average position, and apply the country filter for the United Kingdom. This describes searches that already reach your site, not the whole market, so keep it separate from keyword-tool estimates.

Volume alone is weak evidence. A high-volume term with low commercial intent can define a persona who never buys. Pair volume with the pages that rank and the adverts shown beside them.

Survey and panel data

Panels capture stated preferences. Ask about the last tool the respondent paid for, what triggered that purchase, and who else approved it. Those answers map onto a persona far better than a job title does.

For original survey recruitment, consider YouGov or Prolific. For published audience and market data, check GWI and Statista. Before using any result, check its UK coverage, respondent screening and methodology, and record whether it comes from your own survey or a published dataset.

Sample size matters more than presentation. Thirty responses from one sector cannot support claims about a whole market. State the limit in the document rather than in a footnote.

Review platforms add another source of stated preferences. Read SEO-tool reviews on G2, Capterra and TrustRadius, recording the review date, reviewer role or company details where shown, tool named, use case, and stated pros, cons or switching reasons. Treat these as review evidence, not as a representative survey.

Operational evidence

Your own service data is often the strongest input. The GOV.UK service manual explains how to set performance metrics for your service. The same discipline applies here: name the metric, name the source, set a review date.

Support tickets, trial-to-paid conversion and cancellation notes all describe buyer behaviour. Read fifty cancellation notes and group them by reason. You will often find more reasons than job titles, and they cut across seniority.

In Google Analytics 4 (GA4), review landing-page, session source/medium and conversion-event data. In HubSpot, use contact, company and deal fields such as lifecycle stage, original source, company size, deal stage and closed-lost reason where populated. In Intercom, review conversation tags and attributes, including cancellation or feature reasons where captured.

Market and competitor signals

Market analysis features, such as Semrush market analysis, show how demand moves across a category. That helps you date a persona instead of freezing it.

Moz publishes competitive research resources for comparing rival offers and positioning. Use Moz Link Explorer to compare backlink profiles, Semrush Keyword Gap to find keyword overlaps, or Ahrefs Site Explorer's Organic competitors report to inspect competing search visibility. Then check rival pricing and feature pages to test whether an objection concerns price, migration effort or reporting gaps.

Customer interviews explain the wording behind the numbers. Ask buyers to describe the event that started their search, the alternatives they considered, who approved the purchase and what nearly stopped it; record their words separately from your interpretation.

Turning sources into personas

Write one page per persona. Include the trigger, the constraint, the three sources behind it, and the last review date. Add a line on what would disprove the persona. That line stops it becoming a slogan.

For each export or dataset, record its source, report, collection date, geography and fields. Group queries by intent, then cross-check those groups against survey answers, reviews and conversion or cancellation records. Keep observed behaviour, stated preferences and competitor signals distinct when you write the persona.

Glossary

Common questions

These answers cover how to check the evidence behind a persona and what to do when one source is not enough.

How many sources do I need per persona?

Two at minimum, and they should be independent of each other. A panel answer plus your own cancellation data is stronger than two panel studies.

Can I build personas from keyword data alone?

No. Keyword data shows demand and language, not budgets or approval routes. Add one operational source before you write the persona.

How often should the document be updated?

Review each quarter, or sooner if a major competitor changes pricing. Record the review date on the page so readers know how old the evidence is.

Should personas include demographics?

Only where demographics change the buying route. Company size and sector usually matter more than age or job title.