Building ecommerce SEO buyer personas without invented customer detail

What to take away

A shop selling refillable water bottles from a Leeds unit already holds persona evidence in Search Console, order notes and review text.
Build personas from four source typesyour own query and conversion data, national statistics, platform documentation and regulator guidance on personal data.
State the date and origin of every figure inside the persona. A claim without a source is a guess wearing a job title.
Using named consumer data bought from a data broker can bring data-protection and electronic-marketing duties; check the applicable rules before use.
Personas are working documents. Review them when query mix, returns reasons or ad formats change.

Sources you already own

Use Google's documentation on how to use Search Console to read the Performance report's Search results view. Its Queries and Pages tables show clicks, impressions and click-through rate by query or page. Export the data to Google Sheets, group queries by theme rather than single keyword, and compare the themes with order or conversion data from your store.

Four sources you already own

Export Search Console queries by page
Group queries into themes, not keywords
Add order datapostcode, basket, returns
Mine review text and support emails
Match YouTube ad format to persona

Add order data. Postcode districts, delivery speed chosen, basket value and return reasons describe a real buyer. Review text and support emails supply the language customers use, which feeds page copy and category naming.

National data that anchors personas

Use named ONS datasets rather than citing national data generically: the Census 2021 household composition dataset for England and Wales, Internet users, UK: 2020, and ONS mid-year population estimates by age and sex for local authorities in the UK. They cover household composition, internet use and age structure; check each dataset's geography and reference period before applying it to a persona. Our ONS releases guide provides further context.

England data vs UK data

England-level figures

Use when: Market is England only
Label as: England data
Risk if mixed: Weakens writer brief

UK-wide dataset

Use when: Dataset covers whole UK
Label as: UK data
Risk if mixed: Weakens writer brief

Where your market is England only, use England-level figures. Where a dataset covers the whole UK, label it as UK data. Mixing the two in one table weakens the brief you hand to writers.

Our England audience research guide explains how to sequence the evidence before writing personas. Record the source and reference period in each persona so writers can distinguish measured findings from assumptions.

Data brokers and the law

The ICO's guidance for organisations using marketing services of data brokers explains what a business should check before buying or using such data, including consent and lawful-basis questions. Consider it alongside the UK GDPR, the Data Protection Act 2018 and the Privacy and Electronic Communications Regulations (PECR), as applicable to how the data is obtained and used.

Using personal data from a broker may bring data-protection obligations even if the data is used to shape an audience rather than contact named people. Record the purpose, source and compliance decision; if you plan electronic marketing, check PECR as well.

Aggregate sources, such as ONS tables or your own anonymised order analysis, avoid that problem. For most ecommerce SEO briefs they are enough.

Worked example in pounds

A Sheffield homeware shop turns over £180,000 a year online. Its owner wants three personas and has no research budget.

She exports six months of Search Console queries, groups them into four themes, and finds two drive most non-brand clicks. She pulls England household composition figures from ONS and notes the release date.

She reads the ICO broker guidance and decides against a purchased list, saving a quoted £900 annual licence. Cash cost: £0.

Her time, about two working days at an illustrative £250 a day, is £500 of internal cost.

Each persona then carries a source line: query theme, ONS release or internal order analysis.

Turning evidence into an SEO persona

Use a consistent template in Google Sheets or your persona document:

Persona label and market:a neutral description of the evidence-backed segment and the geography it covers.
Search needs and language:query themes and the wording customers use in reviews or support messages.
Products or categories:only those supported by query, order or conversion evidence.
Evidence:the source, date or analysis period, and what it shows.
Demographic context:a named ONS dataset and its geography, or mark the field unknown.
SEO implications and gaps:the page or content response suggested by the evidence, plus unanswered questions to test.

Worked persona from the shop's available evidence

Label: Non-brand-search-led homeware shopper. Evidence: the Sheffield shop's Search Console export covers six months; its queries are grouped into four themes, and two drive most non-brand clicks. This supports a search-led segment, but the article does not give the query wording, so the specific need, product interest and demographic profile remain unknown rather than being invented.

SEO use: name the actual query themes from the export, then map purchase-intent themes to relevant category or product pages. Use the wording found in queries, reviews and support messages to improve page titles, headings and product information; group informational themes into supporting content clusters linked to relevant pages. Check the search results for each theme for features such as shopping or image results, and choose a suitable page format; record any assumptions that still need testing.

Testing personas against demand

Check each persona against what people actually search. Our article on the ecommerce SEO target audience in England covers how to compare persona assumptions with query and seasonality evidence. That stops a tidy document outliving the market it described.

Set a review trigger. A shift in query mix, a rise in returns for one product group or a new ad format justify revisiting the source lines. If a claim cannot be traced to a dated source, rewrite it.

Common questions

How many buyer personas should an online store have?

Two or three is usually enough to brief writers and shape category pages. More than that duplicates the same evidence under different names, adding maintenance without improving decisions.

Can I build personas without buying research?

Yes. Search Console data, order records, review text and ONS releases cover most of what a small or mid-sized store needs. Purchased consumer data brings the legal duties described in ICO guidance.

What must every persona claim include?

A source and a date. If the figure comes from ONS, name the release. If it comes from your own analytics, name the period. Untraceable claims should be removed.

Do I need separate personas for Scotland, Wales or Northern Ireland?

Only where the evidence differs. If your delivery data or query mix shows distinct behaviour, build it. Otherwise keep one persona and label the underlying data honestly.