A Guide to CSAT Surveys in Customer Support

You can track first response time, resolution time, and tickets per agent down to the microsecond. But here's the hard truth: none of those numbers tell you whether your customers actually felt helped.

A ticket resolved in four hours means nothing if the customer walked away thinking, "That was painful."

This is where CSAT surveys come in. CSAT — Customer Satisfaction Score — is a post-interaction survey that asks customers one straightforward question: "Was the support you just received any good?" No operational gymnastics required. Just direct feedback from the person who experienced your support team.

If you want to know whether your support is actually working, CSAT is how you find out.

What CSAT is and why it works

CSAT measures customer satisfaction with a specific support interaction, expressed as a percentage. The concept is formalised in ISO 10004 on customer satisfaction monitoring — meaning it's not a trendy metric, it's an international standard that's been around for 20+ years.

The survey is sent after a support interaction (typically when a ticket closes), and the customer rates their experience. The format varies, but these are the most common:

1 to 5 stars — Most intuitive. Feels natural to anyone who's rated a film or restaurant.
1 to 5 numeric scale — Where 1 is "very unhappy" and 5 is "very happy." Slightly more formal than stars.
Binary (Good/Bad or Thumbs Up/Down) — The simplest possible format. Highest response rates. No middle ground for fence-sitters.
Emoji scale — Increasingly common because it's visually immediate and skips the whole "numbers" part that confuses some people.

The math is straightforward: you count the number of positive responses (usually 4–5 on a 5-point scale, or "yes" on binary), divide by total responses, multiply by 100, and you have a percentage. A score of 82% means 82% of surveyed customers were satisfied.

The real insight isn't the number itself — it's what the number reveals when you break it down by agent, channel, issue type, or time of day. That's where CSAT becomes genuinely useful.

Why CSAT matters more than you might think

Operational metrics tell you how fast you are. CSAT tells you whether speed matters. A team can respond in 30 minutes and still leave customers frustrated. Another team might take 3 hours and have customers singing their praises. Efficiency and satisfaction are not the same thing. CSAT measures the one that actually predicts whether a customer stays or leaves.

CSAT reveals problems that spreadsheets hide. If your overall CSAT drops, the raw number doesn't tell you why. But if you segment by channel and discover that email satisfaction is 72% while chat is 91%, you've found a specific problem to fix. Same for agent-level segmentation: if one person's average is 76% while the team average is 84%, that's a signal for targeted coaching, not a reason to panic about the whole team.

It drives improvement because measurement creates accountability. Teams that track CSAT consistently outperform those that don't — the UK Customer Satisfaction Index demonstrates year after year that organisations that measure and act on satisfaction see revenue growth correlate with it. The simple act of asking the question makes agents care about the answer.

CSAT often drops before anything else does. Customers get dissatisfied before they churn, before they leave bad reviews, before they switch to a competitor. If you're checking CSAT regularly, you'll catch the problem early enough to fix it.

How to design a CSAT survey that people will actually answer

Keep it absurdly short. One rating question. One optional comment field. Done. Every additional question tanks response rates.

The standard format:

"How would you rate your support experience?" (star rating or binary)
"Anything else you'd like us to know?" (optional)

If you feel the urge to add more questions ("How likely are you to refer us?" "Did our response address the root cause?" "Please tell us what went wrong"), resist. You can always follow up with dissatisfied customers individually. CSAT surveys should take 15 seconds, not three minutes.

Choose a scale that matches your patience for data.

Binary (thumbs up / thumbs down) gives you the highest response rate and the simplest analysis. It's brutal in its simplicity — no "I'm sort of satisfied" nonsense. If you want maximum participation and don't need granular distinctions, binary wins.

A 5-point scale gives you more nuance. You can distinguish between "satisfied" and "very satisfied," or catch the difference between "mildly frustrated" and "extremely frustrated." Slightly lower response rate, but richer data.

A 3-point scale (positive / neutral / negative) is a reasonable middle ground.

For most small support teams, binary or 3-point gets you 90% of the insight with zero extra complexity.

Write a question tied to the interaction, not the company.

Good: "How would you rate your support experience?"
Bad: "How satisfied are you with our company overall?"

The first measures what your team controls. The second is influenced by product quality, pricing, and factors you can't fix with better support. Measure what you can actually do something about.

Make the comment field optional, but give people a place to vent. Comments on negative ratings ("Waited four hours for a response," "Had to explain the issue three times") are gold. They tell you exactly what went wrong. But making the field mandatory tanks response rates, so leave it optional.

When to send the survey

Right after resolution, while the experience is fresh. Send the survey within an hour. If you wait until the next day, response rates plummet and the customer's memory of the interaction fades. The motivation to respond is highest right when the ticket closes.

You can also send it after specific events, not just all resolutions:

After a phone call
When a live chat session ends
After a customer contacts you for the third time on the same issue (to rate that third interaction)
After a VIP or high-value customer's ticket closes

But watch out for survey fatigue. If a customer contacts you weekly, don't send a survey every single week. Set a minimum interval — once per month, or once per customer per month. Over-surveying leads to lower response rates and customers getting annoyed (which defeats the purpose).

What to do with the data once you have it

Collecting CSAT scores and ignoring them is a waste of everyone's time. Acting on them is the whole point.

Review it on a rhythm. Weekly for trends, monthly for deeper analysis. Segment by agent, channel, issue type, customer tier, time of day — look for patterns. A 79% CSAT overall might hide a disaster in your overnight support queue or a specific channel that needs help.

Follow up on every negative rating. When a customer gives you a low score, treat it as an early warning system. Have a manager or senior agent reach out personally. Understand what went wrong. Sometimes you recover the customer and turn them into a loyal one. Always, you learn something valuable about what to fix.

Share the results with your team. Transparency drives improvement. Agents should see their own CSAT scores and understand how they compare to the average. Celebrate positive feedback publicly (it's a powerful motivator). Share constructive criticism privately.

Root-cause analysis isn't optional. When CSAT drops, dig into why. Common culprits:

Long wait times — First response took too long
Multiple transfers — Customer bounced between agents
Wrong information — Answer given was incomplete or incorrect
Tone or dismissiveness — Response felt cold or formulaic
Ticket closed, problem not actually fixed — The classic silent failure

Each one requires different medicine. Long waits need staffing. Transfers need workflow automation or better knowledge bases. Wrong answers need QA processes. Unresolved issues need a "no close until the customer confirms it's fixed" policy.

Benchmarks and what comes after CSAT

90%+ CSAT = Exceptional. You're consistently exceeding expectations.
80–90% CSAT = Good. Solid performance with clear opportunities.
70–80% CSAT = Average. You know what to improve.
Below 70% CSAT = Something's broken. Probably multiple things.

But the absolute number matters less than the trend. An 81% score improving steadily beats an 88% score slowly declining.

CSAT is powerful but not complete alone. Layer it with:

NPS (Net Promoter Score) — Measures overall loyalty ("How likely are you to recommend us?"). Broader than CSAT but less actionable for your support team specifically.

CES (Customer Effort Score) — "How much effort did you have to put in to resolve this issue?" Often predicts future behaviour better than CSAT because it isolates frustration from satisfaction.

First contact resolution rate — Percentage of issues fixed in a single interaction. Strongly correlated with CSAT because customers hate bouncing between agents.

Most modern helpdesk platforms — including Relentify's Surveys product — can automate post-ticket surveys with configurable scales, triggers, and analytics dashboards that segment scores by agent, channel, and category. You can also cross-reference CSAT data with your custom reports and dashboards to spot patterns no single metric reveals.

Common CSAT mistakes (and how to avoid them)

Sending surveys days after the interaction. Response rates collapse. The customer's already moved on.

Over-surveying. Every interaction = every survey = survey fatigue. Pick a reasonable frequency (weekly, monthly) per customer and stick to it.

Collecting data but never acting on it. Asking for feedback you ignore erodes trust faster than not asking at all.

Averaging without segmenting. An 82% overall CSAT might hide a 65% performance in one channel or with one team. Always dig.

Treating CSAT as the only metric. It's one dimension. Pair it with operational metrics like response time and resolution rate, and you get a complete picture.

Frequently Asked Questions

Q: What CSAT score should I aim for?
A: 80%+ is good, 85%+ is very good. But honestly, the trend matters more than the number. If you're at 75% and climbing, that's progress. If you're at 88% and falling, that's a problem. Start measuring and improve from there.

Q: How many responses do I need for CSAT to be meaningful?
A: At least 30–50 per week gives you a reliable signal. Fewer than 10 and you're just guessing. If your response rate is low, consider switching to a binary scale (thumbs up/down) to boost participation.

Q: Should I ask CSAT after every ticket?
A: No. Set a reasonable frequency — once per customer per week or month, depending on their contact volume. Over-surveying tanks response rates and annoys customers.

Q: What if customers never respond to surveys?
A: Low response rates usually mean either the question is too long, the survey arrives too late, or there are too many surveys. Try a shorter form, send it within 30 minutes, and reduce frequency.

Q: How do I know if my CSAT score is actually good?
A: Compare to your own baseline and your industry. SaaS support typically runs 75–85%. Healthcare runs higher (85–92%). Telecom runs lower (65–75%). More important: is your score stable, improving, or declining? That's the real signal.

Q: Can I automate CSAT?
A: Absolutely. Modern helpdesk platforms send surveys automatically after tickets close, collect responses, and feed them into dashboards. Zero manual work required (except acting on the data).

Q: What's the difference between CSAT and NPS?
A: CSAT measures satisfaction with a specific interaction. NPS measures overall loyalty ("Would you recommend us?"). Both are useful. CSAT is actionable for support teams. NPS is more for C-level benchmarking.

Getting started

If you're not measuring CSAT yet, start stupidly simple:

Enable post-ticket surveys (binary scale: good/bad, or 5-point)
Include one optional comment field
Send the survey within an hour of resolution
Review results every Friday
Follow up personally on every low rating
Share results with your team every month

Even this basic setup gives you more insight into how your support is actually performing than any operational metric alone.

From there, you can get fancy — segmentation, root-cause workflows, custom dashboards, NPS and CES layered on top. But start with the basics. Start measuring satisfaction, and everything else follows.