Check local SEO target audience data before you build a campaign

What to take away

Which geographic unit should define the audience?

Start with the boundary, not the persona. A local SEO target audience is defined by where people are, so pick one unit and keep it consistent across every dataset. Postcode districts suit service businesses with tight travel radii. Local authority areas suit multi-town campaigns.

Mismatched boundaries are the most common error. If your keyword research uses towns and your data uses districts, nothing lines up and the audience looks larger or smaller than it is.

Record the choice in the research file. Anyone joining later needs to know why a district was used instead of a county.

What can ONS housing data tell you about an area?

Housing data gives you tenure, household size and dwelling type, which can help frame questions about local service demand. For England and Wales, start with Census 2021 topic tables on tenure, household size and accommodation type in Nomis; use Scotland's Census for Scottish areas and NISRA's Census 2021 results for Northern Ireland. The ONS Postcode Directory (ONSPD) helps match postcodes to official geographic areas. These are area-level context, not counts of a business's prospective customers.

Use it to compare candidate areas, not to describe individuals. A high proportion of flats in one district may support a different service mix and a different landing page structure.

Commercial profiles such as CACI Acorn and Experian Mosaic, and ONS area classifications, can add segmentation context. Treat these as classifications or modelled profiles of places, not observed facts about a client's own customers.

Keep the output short: a table of three to five candidate areas with the measures that matter to your client. More detail rarely changes the decision.

For each candidate area, record the boundary, source and table, measure, reference period, date checked and any caveat. Leave values to be filled from the cited source; do not present an estimate as an observed figure.

Use tools to cross-check signals, not to claim they describe the whole audience. Google Keyword Planner and Google Trends can indicate search interest; Google Business Profile's Performance report shows activity on a client's listings; BrightLocal can track local search visibility. Note the location settings and date, as these outputs are not population counts.

Where do compliance rules limit audience research?

Two regulators shape what you may collect and how you may use it. The ASA AdviceOnline library explains how the CAP Code applies to marketing claims and targeting, including digital activity; check rule 2.1 on identifying marketing communications and rule 3.1 on misleading advertising. The ICO guide to PECR covers cookies, email and tracking used in analytics; in particular, check regulation 6 for storing or accessing information on devices and regulation 22 for direct marketing by electronic mail.

Read both before you install tracking or build an email list from local enquiries. Consent wording and analytics configuration are easier to fix at the start than after a complaint. Before installing non-essential analytics cookies, check PECR regulation 6 and the ICO's guidance; assess email marketing against regulation 22, including its consent and soft opt-in conditions. Under UK GDPR, record the lawful basis under Article 6 and provide the information required by Article 13 when you collect personal data directly. Where you rely on consent for marketing, use a separate, unticked opt-in, for example: “Tick here if you want [business name] to email you about its services.”

If the client sells to consumers, do not assume an enquiry permits follow-up marketing. Apply the PECR rules for the channel and record the basis for processing; distinguish a response to a service request from promotional email.

Which first-party sources beat third-party estimates?

Your own data is the strongest evidence of a real audience. Enquiry logs, call notes and site search terms show what people actually asked for, in their own words.

A structured interview programme turns that raw material into segments. The customer interview method for England explains how to recruit and run those conversations without leading the answers.

Aim for a dozen interviews before you generalise. Below that, treat findings as directional.

How should you record and test the audience?

Write one page per segment: who they are, where they are, what they search for, and what would make them choose the client. Add the source and date for every claim.

Where no source exists, write "assumption" and set a review date. This keeps the campaign honest and makes later testing meaningful.

Then test cheaply. Run a small ad set or a single landing page variant per segment and compare enquiry quality, not just click volume. Update the page when the evidence changes.

For the full sequence from boundary choice to segment testing, the England audience research guide sets out the wider process this article supports.

What should you do when the data disagrees?

Conflicts are normal. ONS averages may suggest one area is stronger while your enquiry log shows another performing better.

Resolve conflicts by recency and specificity. A recent first-party signal usually beats a broad regional average for a single client.

If the conflict persists, split the test. Run both areas for a fixed period and compare cost per enquiry.

Glossary

Common questions

How many areas should I profile?

Three to five is usually enough to compare and choose. More areas add work without changing the decision, unless the client operates nationally.

Do I need interviews if I have ONS data?

Yes. ONS data describes places, not motivations. Interviews explain why people in those places choose one provider over another.

Can I reuse audience research for another client?

Only the method and the public datasets. First-party findings belong to the client whose enquiries produced them, and consent terms may restrict reuse.

How often should the audience be reviewed?

At least once a year, and whenever enquiry quality shifts. Housing and demographic data are updated periodically, so check the source date before relying on it.