CSAT Benchmark Calculator
"Is 85% good?" depends entirely on who's asking. It's top-of-market for an internet provider and unremarkable for a restaurant. Pick your industry, team size, support model and conversation mix, and get a target built for your company instead of a number someone made up.
Learn more about CSAT benchmarksSets the published ACSI benchmark baseline from which all profile adjustments shift.
Lower than most founders assume. Complexity varies enormously by product maturity.
How many agents respond to conversations in your queue.
Whether Fin or another AI agent is answering part of your incoming queue.
How hard the average conversation is to resolve.
Optional: Benchmark Your Recent CSAT
Enter your last period's positive ratings and total ratings. We'll run a Wilson confidence interval test to see if your team is genuinely above or below target.
Target CSAT
Top of market: 90%
A team in SaaS / Software at this scale with this conversation mix should aim for 85%.
CSAT Performance Bands for Your Profile
Roughly where published industry averages sit — no adjustment.
One population of raters, rating one kind of interaction.
The mix industry averages are built on — no adjustment.
| Industry | Average | Strong | Excellent | Notes |
|---|---|---|---|---|
| SaaS / Software | 78–80% | 85%+ | 90%+ | Lower than most founders assume. Complexity varies enormously by product maturity. |
| Retail / E-commerce | 79–82% | 85%+ | 90%+ | Depends heavily on fulfilment and returns — much of your CSAT is decided before support gets involved. |
| Banking / Fintech | 80% | 85%+ | 90%+ | One of the stronger financial sectors. |
| Financial advisors / Wealth | 81% | 86%+ | 91%+ | Relationship-driven interactions score well. |
| Insurance | 76–79% | 82%+ | 88%+ | Claims complexity suppresses satisfaction independently of support quality. |
| Healthcare / Health tech | 77–81% | 83%+ | 88%+ | Hospitals score lower than outpatient care — pick the end of the range that matches your setting. |
| Hospitality / Hotels | 82% | 86%+ | 90%+ | Service expectations are high and so are the scores. |
| Restaurants / Food service | 84% | 88%+ | 92%+ | The highest-rated consumer industry in ACSI. |
| Consulting / Professional services | 84% | 88%+ | 92%+ | High-touch service businesses tend to perform well. |
| Online travel | 75–77% | 82%+ | 88%+ | Travel disruption drags scores down regardless of how the conversation went. |
| Airlines | 74–78% | 82%+ | 88%+ | Operational issues dominate ratings. |
| Utilities / Energy | 74–75% | 80%+ | 86%+ | Customers mostly make contact when something is already wrong. |
| Social media / Consumer apps | 74% | 80%+ | 86%+ | Low baseline trust and moderation disputes hold scores down. |
| Food delivery / Marketplaces | 73–74% | 80%+ | 86%+ | Logistics failures land in the support queue as bad ratings. |
| Internet / Telecom | 68–72% | 75%+ | 82%+ | Historically the lowest-scoring sector. A 78% here is a better result than an 85% in retail. |
The industry averages are published data. They track ACSI industry scores — the most authoritative cross-industry satisfaction source — with the strong and excellent thresholds set above them, consistent with the 64–80% typical help-desk range Zendesk publishes. Ranges appear where sources disagree or a sector splits, like hospitals versus outpatient care.
The adjustments are a model, not data. Nobody publishes CSAT averages cut by team size or conversation complexity. The size, support-model and complexity adjustments are our operating judgement about the three factors that most reliably move a team off its industry average. Each one is shown with its reasoning in the breakdown so you can disagree with it — they're capped at three points each, deliberately, so they nudge the published figure rather than replace it.
There is no official Intercom benchmark. Intercom doesn't publish a platform-wide average, and anyone quoting one at you is guessing. What exists is named customer stories — Bailey Nelson at 95%, 3Commas at 93%, tado° at around 90% through peak season. Treat those as proof of what's achievable, not as a typical result.
ACSI measures whole industries, not support interactions. That's the honest limitation of every benchmark on this page. An industry average tells you roughly where the goalposts are. Your own trend over the last six months, split by team, channel, and AI versus human, tells you whether you're getting better.
- The industry average — this page.
- Your own trend. 83% → 85% → 88% is working. 94% → 91% → 87% is a problem, even though 87% clears the generic bar.
- Human versus AI, tracked separately.
- Team versus team.
- Queue versus queue.
- First response time against CSAT.
- Resolution time against CSAT.
The full argument for each of these is in what a good Intercom CSAT score actually looks like. If your team finds out about bad ratings on Monday morning, route DSAT into Slack as it lands instead.
What is a good CSAT score for my industry?
It varies by roughly 16 percentage points across industries. Full-service restaurants and professional services average around 84%, SaaS around 78–80%, and internet and telecom providers around 68–72%. A 78% score is below par for a retailer and a strong result for an ISP, so compare against your own sector before drawing conclusions.
Does company size affect CSAT benchmarks?
In practice, yes. Very small teams give unusually personal support and tend to score above their industry average, while support organisations of 200+ agents deal with a broader conversation mix, more handoffs, and policy the front line cannot change. This calculator adjusts your target by up to two points in each direction for team size.
Should my AI agent have the same CSAT target as my human team?
No. Intercom reports teammate CSAT and Fin AI Agent CSAT separately and notes that customers rate AI and human support differently, which makes a direct comparison misleading. Set the AI target three to five points below your human target as a starting point, then benchmark the AI against its own historical trend alongside resolution rate and handoff rate.
Is there an official Intercom CSAT benchmark?
No. Intercom does not publish a platform-wide average CSAT score. Published Intercom customer stories report figures between roughly 87% and 95%, and Zendesk puts typical help-desk CSAT at around 64–80%. Those are useful calibration points, not an official benchmark.
How many ratings do I need before my CSAT score means anything?
Roughly 100 ratings gets you a 95% confidence interval narrower than about five percentage points. Below 30 ratings, month-to-month movement is mostly noise. Enter your rating counts in the calculator and it will tell you whether your score is genuinely above or below your target, or whether you simply do not have the sample size to say.