Define what counts as a response
A response rate needs a denominator. A table tent might use scans as the denominator, while a receipt invitation might use the number of receipts handed out. Record the denominator beside the result so the team does not compare scans, visits, and completed surveys as if they were the same thing.
Response rates depend on the whole setup
Audience, placement, timing, question design, trust, and follow-up all affect whether someone responds. A request at checkout has a different context from one on packaging or at a service desk. Compare the same workflow over time instead of borrowing a benchmark from another business.
Put the request next to the customer moment
A code on a receipt, table tent, package, or exit sign should make the question obvious. Match the invitation to the moment: ask about checkout beside checkout, delivery after delivery, or a visit as the customer leaves. Tell people what they will see after scanning and why their answer is useful.
Start with one question
The first screen should ask for a rating or choice that a customer can answer in seconds. A long form raises the effort required to respond and makes it harder to tell which question discouraged completion. Add an optional comment after the first response when the team needs more context to act.
Make scanning feel safe and expected
Customers should be able to see the business name, the purpose of the survey, and a short explanation before they scan. Use a stable destination, a mobile-friendly page, and a sign that says what happens next. Avoid disguising the request as a promotion or asking for information the team does not need.
Use timing to preserve context
Feedback collected close to the experience is easier to connect to a location, team, or interaction. That context can help a team understand an issue while it is still specific enough to investigate, rather than relying on a vague recollection days later.
Measure the useful outcome
Track scans, completed responses, comments, negative alerts, and follow-up separately. A high scan count does not show whether the team learned or acted. Look for the placements and questions that produce enough detail to make a better next decision.