Audience research
Audience research for an online store that starts with the journey, not the recruit
Plan ethical audience research for an England online store, from journey-based recruitment to evidence limits, decision gates and safe follow-up.
Plan audience research for an England store by naming one journey decision, then recruiting people who reached that step inside a stated window. Research agreement stays separate from marketing permission.
What to take away
- Fix one merchant decision and one journey step before writing a screener.
- Recruit on an observable event, not on a demographic label.
- Keep research participation and marketing consent apart, and record your lawful basis separately.
- Version the screener, information notice and discussion guide before fieldwork starts.
- Treat a persona card as dated evidence with a sample count and a retirement trigger.
Name one decision and one qualifying event
Write one research question that changes one merchant decision. Name the seller, journey step, channel and recent period. Then state the observable event that makes a person eligible.
A usable event is already recorded somewhere, such as a delivery-options step loaded during the field window, or a refund policy assessed by an authorised owner. Eligibility is not interest, and being keen on ecommerce qualifies nobody.
Record the eligible population, denominator and period, plus the source system and definition version. A participant statement is not an order, and a small qualitative sample does not describe England. GOV.UK Service Manual planning guidance asks teams to fix the problem, assumptions and next decision before a round opens.
Interview people about the step they actually hit; the store journey interview method opens on a recent event rather than an opinion.
Example: a journey-based screener for a delivery promise
- Decisionwhether to move the delivery cut-off message above the address field.
- Windowstore sessions between 6 and 19 October.
- Eligiblecheckout started, delivery-options step loaded, order not completed.
- Screen outstaff, agency testers, anyone who never reached that step.
- Accessask before booking whether an arrangement is needed.
Keep test prices and delivery terms unambiguous. The CMA's price transparency guidance covers how prices are presented, and research does not create an exception.
Do not ask anyone to enter real card details into an unapproved build. Use test accounts and payment simulations, then set a stop rule.
Keep participation separate from permission
Prepare a participant information notice and version it. The skeleton below covers the controller, the purpose, retention and the withdrawal route.
Controller: [legal entity]. Purpose: research on the delivery-options step. Retention: notes deleted 90 days after sign-off. Withdrawal: email the researcher before reporting.
The ICO's research provisions guidance confirms there is no single lawful basis called research. Our checkout data mapping route sets out what to settle before recruitment opens.
Joining a session does not add anyone to a newsletter. The ICO's guide to PECR covers consent and soft opt-in rules for abandoned baskets.
Before booking, confirm:
Keep participation separate
- Controller and contact point named in writing.
- Purpose limited to the stated decision.
- Retention period and deletion route fixed.
- Withdrawal route explained to the participant.
- Recording and observer arrangements described.
Analyse without promoting a story
Keep four columns during analysis: system observation, participant statement, researcher interpretation and open question. Attach every note to an instrument version, participant code and journey stage.
Report a denominator ladder beside each theme: screened, eligible, invited, enrolled, completed. Missing data is evidence, and people who could not use the method may be the ones the step already fails.
Build persona cards as dated evidence
Use a role label rather than a fictional name and portrait. Include decision context, observed actions and barriers. Add evidence dates, contributing sample, contradictions and a retirement trigger. Do not attach a diagnosis or assumed spending power unless the participant offered it.
Persona cards as dated evidence
- Label
- Recent mobile purchaser using magnification
- Evidence window
- 6 to 19 October
- Contributing sample
- 4 of 11 eligible participants
- Observed barrier
- Delivery cut-off text sat below the address field
- Unknowns
- Whether tablet users behave the same
- Retirement trigger
- Next delivery-promise review, or six months
Retire the card when its trigger arrives, or when the evidence ages out. Store personas as evidence cards shows how to keep the card traceable to the study behind it.
Common questions
Who should I recruit?
People who reached the journey step inside the window, plus the staff who own the fix. A demographic label is not a qualification.
Does research consent cover marketing?
No. Research agreement is a participation control, and marketing email needs its own PECR analysis.
How many participants are enough?
Enough to cover the journey stages and access arrangements you must observe. Report the denominator ladder rather than claiming prevalence.
In this guide
- Defining a store research audience from the decision event outwardsDefine an England store research audience through recent journey evidence, clear roles, safe recruitment, denominators and decision-specific limits.
- Store personas as evidence cards rather than invented charactersBuild evidence-linked store journey personas for England without invented people, stereotypes, false prevalence or unsupported conversion assumptions.
- Running store journey interviews that recruit from a recent event, not an opinionRun safe, reproducible store journey interviews using recent-event eligibility, neutral questions, verified notes and explicit evidence limits.
- Four survey routes for store research, held to one dated evidence gridCompare store journey survey routes through a dated evidence grid covering privacy, security, accessibility, export, deletion, support and cost.
- Ecommerce competitor research: building a dated public claim record without inventing numbersStudy ecommerce competitor pages through a dated claim record that separates public claims, observations and unknowns, then return to customers for buyer evidence.
