Chat-Based Surveys and Feedback Collection

A feedback form emailed after a purchase gets a very low response rate. The same question asked as a WhatsApp message gets many times more replies, because the customer taps one key instead of opening a browser.

The limitation is the reverse: you cannot ask much. One or two questions, and you are done.

Why chat works better

That last row is underrated. In chat, people answer “why?” in their own words far more often than in a form, and those sentences are the useful part.

Keep it to one question

The first question should be the one that matters:

How was your experience today? Reply with a number: 5 — Excellent 4 — Good 3 — Okay 2 — Poor 1 — Very poor

Then one follow-up, conditional on the answer:

• 4 or 5: “Glad to hear it! Would you mind leaving us a Google review? [link]”

• 3: “Thanks — what would have made it better?”

• 1 or 2: “Sorry to hear that. What went wrong? I’ll look into it personally.”

Three questions is the maximum anyone should face in a chat survey.

When to ask

Ask once per interaction. Asking the same customer every week destroys both response rate and goodwill.

Do not survey everyone, every time

• Suppress customers with an open complaint

• Suppress anyone surveyed in the last [30] days

• Suppress customers who opted out

• Sample rather than census if your volume is high

A customer who gets a cheerful “how was your experience?” while waiting for a refund will tell you, loudly and publicly.

Handling the answers

This is the part that matters, and the part usually skipped.

Set a rule: every 1 and 2 gets a human response within 24 hours, with a named owner. Measure how many actually did.

A low score responded to well often produces a more loyal customer than a high score ignored.

Reviews: ask the happy ones

The sequence that works:

1. Ask for feedback privately

2. Promoters are invited to leave a public review, with a direct link

3. Detractors go to your team instead

This is legitimate: you are asking satisfied customers at the right moment. What is not legitimate is paying for reviews, writing them yourself, or offering rewards conditional on a positive review. Those breach platform policies and advertising rules.

See our review response guide.

Other things worth asking in chat

Short, single-question surveys work for more than satisfaction:

• “Which of these should we stock next? Reply A, B or C”

• “Would you prefer morning or evening slots?”

• “What stopped you from ordering this time?” — to enquirers who did not buy

• “What should we add to the menu?”

• “How did you hear about us?”

See our using chat data guide.

Rules and cost

Business-initiated WhatsApp messages need opt-in and approved templates, and category rules affect cost. A post-purchase feedback request is generally treated differently from a marketing message, but confirm current classification and pricing before running these at volume.

See our templates guide.

Watch for bias

• People with extreme experiences reply most

• Staff asking in person get inflated scores

• Low response volumes make month-to-month swings meaningless

• Customers who left silently never respond at all

Treat the trend and the themes as the signal, not a single month’s average.

Close the loop publicly

When feedback leads to a change, say so:

“You told us the evening wait was too long. We’ve added staff from 6 to 9 pm. Thank you for telling us.”

This single habit does more for response rates than any incentive, because customers learn that answering changes something.

Measure

Frequently asked questions

How many questions can we ask in chat? One, plus one conditional follow-up. Three is the maximum.

Should we incentivise responses? Rarely, and never conditional on a positive score. A thank-you works; a reward for a 5-star rating is a policy breach.

What if scores are always high? Check whether staff are asking selectively. Also ask the “what would have made it a 5?” question, which surfaces more than the score does.