All templates

CSAT

Support satisfaction (CSAT)

Three quick questions at the end of a support conversation: how satisfied the customer is, whether the problem is solved, and what to do better.

When to send itIn the reply that closes every support conversation, so the customer can answer while the exchange is fresh.

Free plan, no card needed. “Fit it to my business” rewrites the wording with AI for your brand.

The survey

How did we do?

  1. Page 1
  2. 1How satisfied are you with the help you received? · RequiredCSAT · 1–5
  3. 2Was your problem solved? · RequiredYes · Partly · No, not yetMultiple choice
  4. Page 2
  5. 3What should we have done differently?Be as blunt as you like. It helps us improve.Shown only when question 1 is at most 3 or question 2 isn’t “Yes”Open text
  6. 4Anything the person who helped you did especially well?We pass these on to the team.Shown only when question 1 is at least 4 and question 2 is “Yes”Open text
  • When question 2 is “No, not yet”, go to the “Let's get this sorted” screen.
End screens
  • Thanks for your feedback It helps us give better help next time.
  • Let's get this sorted Sorry it isn't solved yet. Reply to the last email in your conversation and we'll pick it up again.

Suggested coupon

No coupon suggested

A coupon straight after a support chat reads as a bribe for a good rating, and a customer with an open problem wants it solved, not a discount.

Where to send it

  • Support replyPaste the link at the end of the reply that closes a support conversation, in your help desk or email tool. Add ?agent= and ?ref= to record who helped and the ticket number.

When to send it

Support CSAT is a short pulse taken the moment a conversation ends. Put the link in the reply that closes the ticket or email thread, with a line such as "How did we do? It takes 20 seconds." Send it after every closed conversation rather than a sample: support volumes at small companies are modest, and each answer counts.

If your help desk or email tool lets you add details to a link, append ?agent= with the first name of the person who helped and ?ref= with the conversation or ticket number. The customer never sees or types either one. The agent becomes a hidden field you can break results down by, and the reference lets you find the conversation behind a poor rating.

Why these questions

The first question is the classic CSAT question on a five-point scale, from very dissatisfied to very satisfied. Most teams report CSAT as the share of 4 and 5 answers out of all answers. Hearsay's analytics show the full spread, so you can see whether the misses are near-misses or genuinely bad experiences.

Satisfaction and resolution are not the same thing. A friendly agent can leave a customer pleased with the conversation but still stuck, so the second question asks whether the problem was actually solved. Reading the two side by side tells you whether to work on tone, on knowledge, or on the product itself.

On the next page, anyone who rates 3 or lower, or says the problem isn't fully solved, is asked what you should have done differently. Satisfied customers get a different prompt: what the person who helped them did well. Pass those answers on to the team.

When the customer says the problem isn't solved yet, the survey ends on a screen asking them to reply to the last email in the conversation, so the ticket gets reopened instead of disappearing into a report.

Why there is no coupon

A discount offered straight after a support conversation can feel like a bribe for a good score, and a customer whose issue is still open wants it fixed, not money off. Keeping this survey reward-free keeps the ratings honest.

What to do with the answers

Check unsolved answers daily and reopen those conversations first. Each week, look at ratings by agent next to the number of conversations each person handled, and treat a low week as a coaching prompt rather than a verdict: a few hard tickets can pull anyone down. Group the "what should we have done" comments by cause, such as a slow first reply, a wrong answer or a product limitation, and send the product ones to whoever owns the roadmap.

To measure how much work customers had to put in, rather than how they felt about it, try customer effort: support.