Defining a store research audience from the decision event outwards

An audience for store-journey research is not everyone interested in shopping or optimisation. It is a population tied to a recent, observable journey or merchant decision. Define that population before looking at names, profiles or analytics fields.

What to take away

Start with the decision event

Write the decision that the study will inform. A delivery owner may need to decide whether an unclear collection option should be rewritten. A payments owner may need evidence about a failed authorisation route. A product owner may be considering an accessible alternative to an account step.

For each case, specify the seller, channel, journey state, England geography rule and field period. Eligibility could be an authorised record of reaching the relevant step recently, or documented responsibility for the merchant decision.

The Government Service Manual's research planning page recommends linking objectives to a problem, an assumption and the next informed decision. Although written for government services, that structure prevents a merchant from recruiting first and inventing a question later.

Map people to distinct roles

Keep the legal seller, store operator, customer, purchaser, user and other affected person separate. The purchaser may pay for a gift that another person uses. A visitor may appear in an event log without becoming a customer. Neither person can answer for the legal seller.

Inside the merchant, distinguish product ownership, payments and finance, consumer-law review, accessibility, privacy and PECR, product safety and security. Ask participants only about evidence within their experience or authority. A customer describes a recent journey; a payments owner explains documented states; qualified advisers decide legal or assurance questions.

Build the frame from evidence

Record the target population and the actual invitation frame. The frame may omit anonymous abandoners, assisted buyers, people who refused optional tracking or customers using channels outside the selected system. Report those gaps instead of treating the invite list as the whole audience.

The Service Manual says participants should be actual or likely users and discusses recruitment bias in its participant guidance. Use an observable experience criterion. Do not infer residence, disability, protected characteristics, technical skill, vulnerability, purchasing power or organisational authority from a name, postcode, profile, device or transaction.

For a bounded store decision, one practical recruitment channel is a post-purchase invitation on the order confirmation page, aimed at people who completed checkout in the last 14 days. An indicative sample of 8 to 12 participants is typical for one decision.

Ask invitees what participation arrangements they need. The same guidance recommends including disabled participants, while its page on running accessible sessions emphasises individual needs and usable consent. Access requirements belong in session planning, not a stereotype segment.

Publish the denominator ladder

Count records screened, eligible people, people invited, delivered invitations, respondents, enrolled participants, starts, completions and valid answers for each analysed item. Record ineligible, withdrawn and missing cases without exposing identities. These denominators reveal where selection occurred.

The ladder below is illustrative for one bounded store decision.

Agreement to take part does not authorise marketing. Give a participant notice, document the organisation's lawful basis under UK GDPR Article 6 and collect only necessary data.

In the UK, the Information Commissioner's Office publishes guidance on lawful basis and on direct marketing under the Privacy and Electronic Communications Regulations (PECR), whose regulation 22 covers electronic mail. The Service Manual's research privacy guidance recommends early transparency, restricted access and deletion when details are no longer needed.

Decide whether the audience is usable

The product owner should pause if the recent journey cannot be verified, important routes are absent, an assurance role is missing or recruitment would require unapproved tracking. The researcher and statistician should also reject England-wide claims from qualitative or convenience samples. A narrow, honest population can support a bounded store decision; a broad label cannot repair missing evidence.

Named consumer, privacy and PECR, accessibility, product-safety and security reviewers must clear their separate gates before publication or live research. This article supplies a planning method, not a legal conclusion or a finding about England shoppers.

Before you act

Common questions

What should I define first when planning store-journey research?

Write the decision the study will inform. Then set the roles, channel, journey state, England geography rule and field period.

How do I build the invitation frame from evidence?

List the target population and the actual invite list. Note who is missing, such as anonymous abandoners, and use an observable experience criterion. Do not infer characteristics from profile data.

When should the product owner pause the research?

Pause if the recent journey cannot be verified, key routes are absent, an assurance role is missing or tracking is unapproved. Do not make England-wide claims from convenience samples.