Nine questions to ask in technical SEO customer interviews

What to take away

The commonest mistake in technical SEO customer interviews is writing questions after the call is booked, so the session becomes a demo.
Nine question areas cover the buying trigger, the approval chain and the implementation constraint that decides whether recommendations ship, with a final question that widens the sample.
Apply the inclusion criteria firsta buyer or blocker in an organisation with a crawlable site, not a competitor's employee.
Treat answers as qualitative and pair them with separate demand data.
Use one question set every time so answers compare across roles.

Why interviews beat survey data on their own

A survey tells you what a population reports. It rarely explains why a recommendation was parked for two quarters. Interviews fill that gap when you ask about the last decision.

Recruitment differs by market, but the questions do not. Our audience research guide for England covers sampling and how findings feed a strategy document.

The inclusion criteria to apply

Recruit people who bought, blocked or implemented technical SEO work in the last 18 months. One participant per organisation suits a first round.

Inclusion criteria for interviews

Bought, blocked or implemented work in 18 months
One participant per organisation in first round
Exclude sellers of the same service
Exclude free-tool users without budget authority
Note sector, site size and platform

Exclude anyone selling the same service, and free-tool users with no budget authority. Note sector, site size and platform so answers group later. For example — and these figures are illustrative — a team paying £400 a month for a managed crawl platform sits in a different bucket from an in-house team paying £3,000 a month for technical SEO audit tools and the staff to run them.

Nine questions for technical SEO customer interviews

Walk me through the last technical SEO problem you paid to solve. Anchor the session on a real decision.
What triggered the search? Look for a migration, a traffic drop or a new hire.
Who else had to agree before money moved? This maps the approval chain.
What did you compare before choosing? Ask about the crawl and audit tools on the shortlist — Screaming Frog, Sitebulb or Semrush — because you are recording a process, not endorsing a product.
Which output did you actually use? Ask for the artefacta crawl export, a structured data report or a prioritised backlog.
What stopped you acting on the last recommendation? Engineering capacity and release freezes usually surface here.
How did you check the work was correct? Validation habits show which evidence your deliverables need.
What would have made the decision easier? Answers often point at pricing clarity or a pilot.
Who else should I speak to? Snowball sampling is legitimate if you record how each participant was recruited.

What to read alongside the answers

Interviews describe motivation; they do not size a market, so read them alongside separate demand data.

Before publishing claims built on these sessions, check they would survive a regulator's eye. The buyer personas data and sources guide on this site sets the evidence standard for persona work.

Where a participant describes validating markup, compare that against the catalogue of structured data types Google Search supports — Article, Product, FAQ, Breadcrumb, LocalBusiness and Event among them — so follow-up questions stay accurate.

If the organisation is a digital service with formal governance, the technology guidance in the GOV.UK service manual shows how those teams describe assurance.

Where a participant raises information governance in a regulated sector, the standards catalogue from BSI is where to check whether a standard applies; BS ISO/IEC 27001 for information security management is the usual starting point.

How to run the sessions

Keep each interview to 30 minutes and ask one question at a time. Record only with written consent and a lawful basis recorded under the UK GDPR, following ICO guidance.

Do not pitch. If the participant asks what you sell, answer briefly and return to the script. Write a one-page summary within 24 hours.

After six to nine sessions, group answers by trigger, blocker and validation method, which makes the research usable rather than anecdotal.

Common questions

Short answers to what comes up most when teams plan the sessions.

How many interviews are enough?

Six to nine sessions usually surface the recurring triggers and blockers in one segment. Add sessions when a new role appears.

Should I pay participants?

A modest incentive improves attendance among busy technical staff. Check your organisation's policy and record the amount.

Can I reuse the same questions for agencies?

Yes. Change the framing from buyer to seller but keep the sequence so both sets compare.

What if the participant cannot name a budget?

Ask who signs off the invoice instead. The approval route matters more than the figure.