How to Get Freshdesk CSAT Surveys to Actually Get Filled In
Freshdesk CSAT surveys ask one question after a ticket is resolved and record a 4-point rating. Response rates above 20% are achievable; most teams sit under 10% because the survey fires immediately on solve, before the customer has tested the fix.
The survey itself is free to configure. Satisfaction reports, though, sit on Freshdesk Growth and above — on the free plan you can collect ratings and barely see them.
That mismatch is worth knowing before you spend a week on survey design.
How does Freshdesk CSAT actually work?
When an agent resolves a ticket, Freshdesk can send an email with a question and a rating scale. The response is stored on the ticket and aggregated into a satisfaction score. The default scale has four faces — very unhappy to very happy — and by convention anything at the top two counts as positive.
Configure it under Admin > Productivity > Email notifications or the CSAT section, depending on your tenant version. You can edit the question text and add an optional comment box.
Step 1: Delay the survey instead of firing on solve
The default sends on resolve. Change the trigger to send a few hours later — 3 to 4 hours is a good starting point for email support, longer for anything with a shipping or deployment step.
Why: on resolve, the customer has read a message but hasn’t confirmed the fix worked. They rate the interaction, not the outcome. A few hours of delay means they’ve either used the product successfully or hit the same bug again, and both produce a more honest score.
Step 2: Keep one question and make it concrete
The default “How would you rate the support you received?” gets answered on mood. A better wording: “Was your issue resolved?” with the same faces. It’s answerable, and it produces data you can act on.
Add the comment box but don’t require it. Required comments drop response rates hard. The comments you do get will skew negative — that’s normal and still useful.
Step 3: Skip the survey where it’s noise
Not every ticket deserves a survey. Configure the CSAT to skip:
- Tickets resolved automatically by a rule with no human reply
- Tickets tagged
spamorinternal - Tickets resolved within 10 minutes of creation (usually simple how-tos)
Every survey sent to a ticket nobody worked on lowers your average and tells you nothing.
Step 4: Read the score correctly
A 85% satisfaction rate on 40 responses a month has a wide error margin. Treating month-to-month swings as signal produces pointless panic. Look at rolling three-month windows, and look at the response count alongside the percentage.
Low response volume is itself the finding. If you send 400 surveys and get 18 responses, the sample is unrepresentative — usually unhappy customers are over-represented, because happy ones don’t bother.
What’s the realistic plan requirement?
| Need | Free | Growth |
|---|---|---|
| Send CSAT survey | Yes | Yes |
| Satisfaction dashboard | No | Yes |
| Filter by agent/group | No | Yes |
| Export responses | No | Yes |
If CSAT matters to your reporting, you’re on Growth ($15-19/agent/month) or above. Collecting data you can’t read is a common and avoidable trap on the free tier.
Does a bad CSAT score mean bad support?
Often it means slow first response, not bad agents. Split the data by first-response time before drawing conclusions. Teams I’ve looked at usually find that tickets answered within an hour score double those answered the next day, regardless of whether the answer was good.
That reframes the fix: it’s a staffing and routing problem, not a training problem.
When should you not run CSAT at all?
If you’re a two-person team with ten tickets a week, talk to your customers directly instead. A survey at that volume adds process and produces statistically meaningless numbers. Also skip CSAT if you’re not going to act on it — asking for feedback and ignoring it costs more goodwill than never asking.
Quick Summary
- Delay CSAT by 3-4 hours after resolve; the default timing inflates neutral scores.
- Ask “Was your issue resolved?” rather than rating the interaction.
- Exclude auto-resolved and spam tickets or your average is polluted.
- Satisfaction reports need Growth; on Free you collect data you can’t analyse.
- Judge scores on rolling three-month windows and response counts.
