What Is a Good CSAT Score? How to Calculate It, How to Read It and a 30-Day Plan to Lift It

What is a good CSAT score is a question with no single number as the answer, because the score depends on how you ask, who answers and what you sell. This guide shows how CSAT is calculated, why your own trend matters more than any benchmark, how to read a thumbs-based score, and a 30-day plan to improve it.

90%example: 45 of 50 happy ratings (illustrative)
1formula: happy ÷ all ratings × 100
30days in the improvement plan

⚡ In short

  • CSAT is the percentage of ratings that are positive. Happy ratings divided by all ratings, times 100.
  • There is no universal pass mark. Many teams treat the high 70s to 80s as healthy, but this varies by industry and by how you ask. Treat any benchmark as a rough guide.
  • Compare yourself with your own last month first. A steady rise matters more than a single number.
  • Check the sample size. Ten ratings can swing wildly; a few hundred are far steadier.
  • Read the comments. They tell you what to fix; the percentage only tells you whether to look.
What is a good csat score in LiveChatBot: the chat widget on an online store answering a delivery question
How it looks in LiveChatBot: the chat widget on an online store answering a delivery question.

Definition

What CSAT means and how it is calculated

CSAT stands for customer satisfaction score. You ask a customer how satisfied they were with a specific interaction, such as a chat, and count the positive answers.

The formula

CSAT % = positive ratings ÷ total ratings × 100

If 45 customers out of 50 who rated a chat said it was good, your CSAT is 90%. That example is illustrative. The same method works whether you use a five-point scale, where 4 and 5 count as positive, or a simple thumbs up and thumbs down, where a thumbs up counts as positive.

What CSAT is not

CSAT measures one moment, not the whole relationship. It is different from NPS, which asks how likely someone is to recommend you, and from a repeat-purchase rate. A customer can be happy with a chat and still never buy again.

Benchmarks

So, what is a good CSAT score?

Here is the honest answer: it depends. Industry reports quote very different averages because each one uses a different scale, a different way of asking and different customers.

A rough rule of thumb

Many support teams treat a score in the high 70s to the 80s, as a percentage, as reasonably healthy, and above 90% as excellent. Treat that as a rule of thumb, not a measured fact, and do not copy a number from a blog and call it your target.

Why benchmarks mislead

  • Different scales. A five-point average and a thumbs-up percentage are not the same measure.
  • Different timing. Asking right after a reply gets different answers from an email sent a week later.
  • Selection. Very happy and very unhappy people answer most. The quiet middle does not.
  • Different questions. Delivery, product and support get mixed up in scores.

The most useful target is your own: beat last month, keep the lows low and learn from every thumbs-down.

Sample size

What is a good CSAT score on only a handful of ratings?

A percentage looks precise, but small samples are noisy.

Imagine you had 4 ratings this week, and 3 were positive. Your CSAT is 75%. One more unhappy customer next week and it drops to 60%. Nothing about your service changed; you just had few ratings.

A practical way to handle this

  • Report CSAT by month, not by day.
  • Always show the number of ratings next to the percentage, such as 88% from 120 ratings.
  • Wait until you have a decent count, such as 50 or more, before drawing conclusions.
  • Look at the trend across several months.

The response rate

The share of chats that get rated matters too. If only 5% of chats get a rating, the score reflects a small and perhaps unusual group. A one-tap rating tends to get a better response than a long form, which is why simple thumbs work well for small teams.

Thumbs

What is a good CSAT score when you use thumbs, not stars?

If your tool gives only a thumbs up or thumbs down, the method is the same, with a simple twist: the score is thumbs up divided by all ratings.

LiveChatBot works like this. After someone on your team replies in a chat, the visitor can rate it with a thumbs up or down and add a comment, if you have turned it on in Settings, Widget design, Features. Reports show a Satisfaction % and list the ratings and comments under Feedback. You can read the chat satisfaction survey page for the full set-up.

What to remember

  • A thumbs score is a bit more generous than a 1 to 5 score, because a mildly pleased visitor will often tap a thumbs up where they might give a 3 or 4 on a scale.
  • It is not a 1 to 5 CSAT or NPS survey, and there are no emailed surveys, so do not compare it directly with scores from other tools.
  • Use it to track your own trend and spot weak chats.

30-day plan

A 30-day plan to lift your CSAT

Small, steady changes beat one big overhaul.

Week 1Measure the baseline

Turn on ratings, collect a week of data and note your Satisfaction % and the number of ratings. Read every thumbs-down comment.

Week 2Fix the top three complaints

Group the comments. Usually a few themes repeat: slow replies, wrong facts or no handover to a person. Fix the biggest three.

Week 3Speed up and standardise

Write [saved replies](how-to-use-saved-replies-in-live-chat) for the repeat questions, check your opening hours and use the away form after hours.

Week 4Prevent repeat questions

Turn the top questions into help articles, so customers find answers before they chat. Compare this week’s score with week 1.

Worked example

A worked example: reading one month of ratings

Here is an illustrative month for a small online shop (numbers made up).

  • Week 1: 20 ratings, 15 positive, which is 75%.
  • Week 2: 25 ratings, 19 positive, which is 76%.
  • Week 3: 30 ratings, 25 positive, which is 83%.
  • Week 4: 25 ratings, 22 positive, which is 88%.

For the month, that is 100 ratings and 81 positive, so CSAT is 81%.

What the owner learns

The owner read the thumbs-down comments in week 1. Most said “slow reply” or “wrong delivery date”. She wrote saved replies with the correct delivery times, fixed her opening hours and switched on Android alerts. The score rose as replies became faster and more accurate.

What she does not conclude

She does not claim 81% is good or bad against other shops. She notes that it is higher than week 1, has a decent sample, and that the comments now mention different problems. Next month she tries to beat 81%. That is the sensible way to answer what is a good CSAT score for your own business.

Quick wins

Quick wins that usually move the score

Most small teams find the same few levers work.

Reply faster

Waiting is the biggest source of unhappy ratings in many teams. Use alerts on the Android app so replies are quick, and see live chat response time for how to measure and improve it.

Get the facts right first time

A confident wrong answer is worse than a slow right one. Keep prices, hours and policies current in your saved replies and, if you use the AI Agent, in its knowledge.

Hand over early

If a visitor asks for a person, hand over quickly. A visitor asking for a human triggers a red alert for your team in LiveChatBot.

Close the loop

If you cannot solve it in the chat, make a ticket and tell the customer when they will hear back. Tell them what you did, not only what you will do.

Using LiveChatBot

Track the score in LiveChatBot without extra tools

You can do the whole routine inside the dashboard. Switch on ratings, watch the rating badges in the inbox, then open Reports to see the Satisfaction % and the comments. The by-agent table helps you see conversations and replies per person, and the Reports also show first response time and response rate, which often explain a low score.

LiveChatBot is free to start, with two websites on the free plan, 100 free visitors per website per month and no per-seat charges. Details are on the pricing page. If your customers keep asking the same things, the help center software page explains how articles cut repeat questions.

What it will not do

It will not calculate NPS, send emailed surveys or score the sentiment of comments. You read the comments yourself.

Frequently asked questions

What is a good CSAT score for a small business?

There is no universal number. Many teams consider the high 70s to the 80s healthy and 90% or more excellent, but compare yourself with your own trend first.

How do I calculate CSAT from thumbs ratings?

Divide thumbs up by all ratings and multiply by 100. For example, 45 out of 50 is 90% (illustrative).

How many ratings do I need before the score means anything?

As a rough guide, 50 or more per month. With fewer, one unhappy customer can move the percentage a lot.

Is CSAT the same as NPS?

No. CSAT asks how satisfied someone was with a specific interaction. NPS asks how likely they are to recommend you.

Does LiveChatBot show a Satisfaction score?

Yes. Visitors can rate a chat with a thumbs up or down and a comment. Reports show Satisfaction % and the Feedback list.

Keep reading

Related guides

📘

Chat Satisfaction Survey: Ask Visitors to Rate Every Chat With a Thumb and a Comment, Then See the Score in Reports

Read the guide →

🧭

How to Ask for Feedback After a Chat: When to Ask, What to Say and How to Use the Answers

Read the guide →

Start measuring satisfaction

Switch on chat ratings and watch your first Satisfaction % this week. Start free on livechatbot.in.