Audience research
Part of How to run shopping feeds audience research for multichannel shops
Five steps for using customer interviews to plan shopping feed changes
A five-step method for using customer interviews to plan shopping feed changes, from recruiting shoppers and consent to running sessions and tagging findings.
What to take away
- Decide whether interviews will beat the sales and returns data you already hold.
- Recruit shoppers who bought in the last 90 days, plus basket abandoners.
- Write six to ten open questions tied to feed attributes, not to your brand.
- Record consent before you record anything, within what ICO guidance allows.
- Give every finding an owner, a feed field and a channel.
Step 1: Set the inclusion criteria before you recruit
Decide who counts as a customer
You need a defensible rule for who is in scope. A workable one is anyone who bought in the last 90 days, plus anyone who abandoned a basket in the last 30 days. Exclude staff, trade accounts and anyone who has asked not to be contacted.
Inclusion criteria for customer interviews
- Bought in last 90 days
- Abandoned basket in last 30 days
- Exclude staff and trade accounts
- Exclude do-not-contact requests
- Recruit returners for attribute gaps
- Recruit in Scotland, Wales, Northern Ireland
Mix the sample. Two returners will tell you more about attribute gaps than five happy buyers, because a return usually follows a mismatch between the listing and the item.
Geography matters. Consumer protection rules apply across the United Kingdom. If your returns or delivery terms differ in Scotland, Wales or Northern Ireland, recruit there too.
Agree what you will not ask
Interviews drift into pricing opinions and design taste. Park those and keep the scope on attributes. The audience research guide for England sets out the wider research plan this listicle sits inside, so read it before you settle on a method.
Step 2: Write questions that map to feed fields
Draft six to ten open questions
Open questions produce usable quotes; closed ones produce yes and no. Ask a shopper to describe the last time they compared two similar items and what information they wanted at that moment.
Writing questions that map to feed fields
- Draft six to ten open questions
- Ask about last comparison moment
- Probe colour, size, material, compatibility, delivery
- Order questions broad to narrow
- Avoid leading phrasing
- Pilot with two people first
Then probe the fields you control. Colour, size, material, compatibility and delivery window are all attributes you submit to a channel. Google's guidance on submitting product attributes and attribute values explains how those fields are structured, so phrase questions around attributes you can change.
Order the questions from broad to narrow. Start with the shopping moment, then narrow to the fields. Avoid leading phrasing, such as asking whether the size guide was clear, because that invites agreement.
Pilot with two people first
Run the script past two colleagues or two willing customers. If a question produces a blank look, cut it. Ten questions is a ceiling, not a target.
Step 3: Recruit, consent and record properly
Follow the direct marketing rules
Interviewing existing customers is research, not marketing, but the boundary blurs once you ask permission to send offers afterwards. The ICO's step-by-step guide to direct marketing for your small business explains consent and lawful basis in plain terms, so read it before you draft your recruitment email.
If you quote a customer in an advert, the UK Code of Non-broadcast Advertising and Direct & Promotional Marketing sets out what you may claim and the evidence you need.
Keep the admin small
One spreadsheet, one consent column, one folder of notes. Anonymise as you write the notes, and delete recordings once they are checked. Store consent records separately from the notes, and note a deletion date next to each recording.
Step 4: Run the sessions and capture evidence
Structure each session
Allow 30 minutes: five minutes of context, twenty minutes of questions, five minutes to confirm what you heard. Ask for a specific example whenever you get a general statement.
Watch the path, do not just listen. Screen sharing during a live search shows which filter a shopper reaches for first, which is evidence a script alone will not give you.
Log findings against feed fields
Tag each quote with the field it touches: title, image, size, colour, delivery. A short tag list beats a transcript you never reread.
Step 5: Turn findings into feed changes
Write one change per finding
A finding without an owner is a story. Give each change a field, a channel and a person. If three shoppers say the size guide is unclear, that becomes a change to the size attribute and the size chart, not a slide.
The survey tools and data sources article covers how to size and track that backlog across channels.
Re-interview after one quarter
Feeds, catalogues and channel requirements all change. Repeat with a fresh group rather than the same six people.
Common questions
How many interviews do I need?
Six to twelve is a practical working range for a small retailer. Stop when new sessions repeat what you have already heard.
Can I interview customers outside England?
Yes. Data protection and consumer rules apply across the UK, but if your delivery, returns or pricing differs in Scotland, Wales or Northern Ireland, recruit there too.
Do I need consent to record a call?
Yes. Say what you are recording, why, and how long you will keep it, then get agreement before you start.
What if I cannot recruit anyone?
Use order notes, returns reasons and support tickets instead. Interviews are one method, not the only one.
