Audience research
Desk data or real people for keyword research audience research?
A practical guide to keyword research audience research in England, covering ONS data, surveys, interviews, competitor gaps and local targeting, plus a glossary.
What to take away
- Audience research answers who is searching and why, before a keyword tool reports how often.
- Desk data firstthe ONS bulletin Internet access, households and individuals, Great Britain reports internet access in 96% of households in Great Britain.
- Six to ten interviews plus a five-question survey supply the wording.
- Every cluster becomes one page with one job.
- Local intent needs official place names, not invented districts.
What desk data can and cannot tell you
Desk research uses data somebody else collected. It is quick, cheap and a sensible first pass.
The ONS publishes its bulletin Internet access, households and individuals, Great Britain each August. The latest edition reports internet access in 96% of households in Great Britain.
Platform data covers a different slice. LinkedIn sizes a professional audience by job title, industry and seniority. Statista's TikTok hub describes a younger base.
Desk research sources
ONS internet access
- What it shows
- How people use internet
- Best for
- Search reliance
- Example
- Frequency questions
- Limitation
- National, not local
Platform ad data
- What it shows
- Job title and industry
- Best for
- Decision-maker buyers
- Example
- LinkedIn targeting tools
- Limitation
- Platform-specific users
Published figures describe groups, not your buyers. Build a rough profile, then test it on real people.
Example: a Manchester cycle shop
The shop sold commuter bikes, servicing and accessories, and its pages repeated its own product names. A survey and eight interviews produced three clusters.
Manchester cycle shop
| Customer wording | Cluster | Page and its one job |
|---|---|---|
| "a bike for riding to work" | commuter bikes Manchester | Buying guide for a five mile commute |
| "somewhere that services what it sells" | cycle servicing Manchester | Servicing page with turnaround times |
| "Saturday opening matters" | cycle shop opening hours | Location page with hours and parking |
Each cluster became a brief naming the segment, the quote and the primary phrase, so the buying guide could not drift into servicing copy.
Surveys: five questions that return language
Surveys give reach. Two hundred customers answering five questions show patterns within a week.
Surveys: five questions
- What did you type into Google the last time you needed this service?
- How would you describe the problem to a friend?
- How did you compare two suppliers?
- What nearly stopped you buying?
- What would you search for if your usual supplier were closed?
Keep one open text field per question, because free text is where a keyword list starts. Pilot the questions on three people.
Panel suppliers charge per completed response, so ask for a quote. This comparison of keyword research survey tools in England explains why panel access matters more than the subscription price.
Interviews: six to ten conversations
- Recruit ten customers from your own list, not friends.
- Ask about the last time they needed what you sell, never about your website.
- Let silences run; the useful phrase often arrives after a pause.
Start with "walk me through the last time you needed this". Then ask what they typed into a search engine, and what they expected to find.
"I did not want a sales call. I just wanted the price and to know if they could come out on a Saturday."
Highlight every phrase that sounds like a search query and keep the quote beside it.
Recording answers is processing personal data, and the ICO guide to the GDPR covers consent and retention.
Competitor pages as audience evidence
- Shortlist five competitors ranking for your main phrase.
- Record each page's title, headings and the query it answers.
- Read reviews, forum threads and comments for customer wording.
- Note the wording their adverts use.
- List topics none of them covers.
Gaps are the opportunity. If every rival writes for London buyers, the commuter towns around a city may be open. A structured method for turning rival pages into a shortlist sits in keyword research competitor research in England.
Local research for an English service area
Local intent adds place names, landmarks and "near me". List the towns, boroughs and postcode districts you genuinely cover, then check whether your pages name them.
Use the ONS UK geographies methodology for standard English region, county and district names rather than guessing.
Residents use shorthand too, such as a shopping centre, a road or a district nickname. If people say "the Arndale" rather than "city centre", that phrase belongs in a heading.
Moz Local and Semrush handle listings and local rank tracking, but do not build a thin page for every village.
Turning findings into a keyword and content plan
- Group every quote and phrase into themes; each theme is a cluster.
- Tag each phrase with a stage of needlearning, comparing or buying.
- Check volume and difficulty, but let interview evidence break ties.
- Give each cluster one page with one job.
- Write the brief with the segment, the quote and the phrase list.
- Review quarterly, because search language moves.
Your own query data is the cheapest evidence you have. The Google Search Console performance report lists the queries your pages already appear for.
Grouping phrases so one page can serve them is the step most teams rush. A step-by-step method sits in keyword clustering for UK content planning.
From research to plan
- Group findings by theme
- Build clusters around core topics
- Map phrases to stage of need
- Prioritise by evidence, not volume
- Write audience profile into each brief
- Set quarterly review rhythm
- Give each page one job
Time and money before you start
A five-question survey takes an hour to draft and a week to field. Ten interviews take about two days with recruitment and write-up. Panels bill per completed response, so ask for a quote first.
Common questions
How many interviews do I need?
Six to ten for a small business. The same phrases usually repeat by the seventh or eighth conversation.
Can I skip surveys if I already have interviews?
You can, but surveys show how common a phrase is. Interviews explain why people search that way.
Do I need consent to record a customer interview?
Yes. Say how you will store the recording and get permission before you start.
How do I turn quotes into a keyword list?
Highlight phrases that sound like search queries, group them into clusters, then check search volume.
In this guide
- When to build a keyword research target audience from search dataUse keyword research to spot recurring needs, check them against audience evidence and decide when a segment is ready for a content plan.
- Buyer personas from keyword research or from interviews?Contrasts persona building from search keyword data with interview-led research, showing how England teams blend both to plan content that matches real local demand.
- Do keyword research customer interviews beat survey data?A buyer's guide to choosing between customer interviews and surveys for keyword research, with inclusion criteria and a decision table for England.
- Why keyword research survey tools hinge on panel access, not subscriptionsCompare keyword research survey tools for England teams: panel samples, DIY forms and hybrid stacks, with costs, compliance points and a worked example.
- How to run keyword research competitor research for an England audienceA checklist for keyword research competitor research in England: choosing rivals, finding keyword gaps, filtering by intent and measuring positions.
