Social commerce customer interviews: six prompts that map a shopping journey

The best interview questions reconstruct one recent journey. They do not ask people to predict whether they would buy from social media in general. The six prompts below cover discovery through follow-up, with each included because an authoritative record identifies a decision or evidence boundary at that stage.

Recruit people linked to a defined event, such as a completed order, abandoned checkout or return delivered in England. Do not describe the sample as representative of England unless its design supports that claim.

Use a neutral invitation that names the research purpose, duration, incentive and data use. An interviewer has no responsibility for the participant's refund or complaint outcome.

At the start, confirm the specific order or browsing event without disclosing information the participant has not supplied. Ask permission before recording, and offer a genuine route to continue without it if the protocol allows notes.

What to take away

1. "What were you doing when you first noticed this product?"

Let the participant explain the setting, content and device without supplying a platform name. Ofcom's Online Nation 2025 shows that UK online behaviour is distributed across services and devices. Its reach data does not prove where an individual purchase began, which is why unaided recall comes first.

2. "What did you do before opening the product or shop?"

This reveals whether comments, a creator, search or another account supplied proof. The ASA's influencer-recognition study found that confidence did not consistently translate into accurate recognition of influencer adverts in its UK research, focused mainly on TikTok and Instagram. Follow by asking what the participant believed about payment, gifts or commission.

3. "Which product claim did you need to check, and where did you check it?"

Ask for the exact claim, not whether the post felt trustworthy. The CMA's online choice architecture review examines how choice information and interface design can shape decisions. It does not prove harm in this journey, but it supports inspecting where evidence and paid placement appeared.

4. "Before paying, what did you understand about the seller, delivery and returns?"

The GOV.UK distance-selling guide lists information required for covered UK distance contracts and outlines delivery and cancellation duties, with exceptions. The interview can show what a buyer found and understood. A separate legal review must determine whether the actual journey complied.

5. "What role did ratings, reviews or endorsements play?"

Ask which review was used, who appeared to provide it and whether an incentive was visible. The CMA's fake-reviews guidance covers businesses that publish reviews or information derived from them. Do not combine an independent customer review, an incentivised review and a paid creator post into one theme.

6. "What happened after the order, including any messages you received?"

This uncovers confirmation, support, delivery, cancellation and promotional follow-up. The ICO's electronic-mail guidance addresses direct marketing through channels including email and text. Whether a message counts as marketing depends on its content and context, so preserve examples for review rather than accepting the business label.

Analyse answers without inventing a majority

Code each journey by observed stage, unresolved question and outcome, keeping conflicting cases. Six interviews can reveal a failure mode worth testing, but cannot establish a national percentage.

Report recruitment source, dates, eligibility, England rule and incentives alongside the themes. The most useful output is a specific change and a measure: moving verified delivery information before the hand-off and tracking retained orders, not a vague claim about trust.

After each session, separate the participant's account from the researcher's interpretation, compare it with permitted order or support records, and mark any mismatch without deciding one source is correct.

Stop recruitment when planned groups are covered, not merely when later conversations sound familiar. A different checkout route or returned product can expose a decision the earlier sample never faced.

Before you act

Common questions

Why should interviewers avoid naming a platform?

The article says to let the participant explain the setting, content and device without supplying a platform name. Ofcom's Online Nation 2025 shows UK online behaviour is distributed across services and devices, but its reach data does not prove where an individual purchase began, so unaided recall comes first.

What should be done with conflicting accounts?

The article says to keep conflicting cases and compare the participant's account with permitted order or support records. Mark any mismatch without deciding that one source must be correct. Stop recruitment when the planned groups are covered, not merely when later conversations sound familiar.

What makes a useful research output?

The article says the most useful output is a specific change and a measure, such as moving verified delivery information before the hand-off and tracking retained orders. It should not be a vague claim that shoppers want trust, and six interviews cannot establish a national percentage.