CSAT in BPO stands for customer satisfaction score: the rating a customer gives right after a call, chat or email with an outsourced support agent. It is usually the share of survey responses in the top ratings, such as 4 or 5 on a five-point scale, shown as a percentage. In a BPO it sits on every agent scorecard next to handle time and quality, and it is often a contracted target. To improve it, fix the causes customers name, coach agents on scored contacts and close the feedback loop.
That is the short answer. This guide covers how CSAT is measured and calculated, how outsourcing partners use it, what a good score looks like, and a nine-step CSAT improvement program. If you want to see how an outsourced team scores and coaches its agents, our quality assurance page sets out the process.

What is CSAT Score?
CSAT stands for customer satisfaction. It is a number that shows how happy customers are after talking to your support team, whether that is your own team or a call center you have hired. Call centers and BPOs track it on almost every contact.
Here is how it is measured:
- The customer contacts your business or your outsourcing partner by phone, chat, email or messaging.
- When the conversation ends, they receive a short survey.
- The survey asks one main question, such as “How satisfied were you with your experience today?”
- The customer picks a number, often on a scale of 1 to 5: – 5 means very satisfied – 4 means satisfied – 3 means neutral – 2 means dissatisfied – 1 means very dissatisfied
If most answers are 4s and 5s, your support team or your call center is doing its job. If you see more 1s and 2s, customers are unhappy with something, and the usual causes are slow replies, long hold times, being passed between agents, a problem that was not solved, or an agent who did not listen.
CSAT measures one interaction. That is what makes it useful in a BPO: every response can be tied to a specific agent, queue, channel and reason for contact.
How to calculate CSAT
The most common method counts the satisfied answers:
CSAT (%) = (number of 4 and 5 responses / total responses) x 100
For example, if 400 customers answer the survey in a week and 340 of them choose 4 or 5, your CSAT is 340 / 400 x 100 = 85%.
This is also how the main help desk tools treat the scale. Zendesk’s CSAT documentation maps survey answers to good and bad: on a 1 to 5 scale, 1 to 3 count as bad and 4 to 5 as good. On a two-point scale, 2 is good.
Some teams use an average instead. Call Centre Helper’s guide to calculating CSAT sets out three versions: the sum of all scores divided by the maximum possible, happy versus unhappy buttons, and star ratings. Whichever you pick, use the same one every month and in every contract, or the trend means nothing.
How CSAT works in a BPO
In an outsourced setup, CSAT is not only a report. It is how the brand and the partner agree what good service looks like.
- It is attributed to agents. Each response is linked to the agent who handled the contact, so CSAT becomes a personal score as well as a program score.
- It sits on the scorecard. Agents are usually measured on CSAT alongside quality assurance scores, average handle time and first contact resolution.
- It is broken down by driver. A good BPO splits CSAT by channel, contact reason and queue, so a drop can be traced to a policy, a product issue or a coaching need.
- It is often in the contract. Many brands set a target CSAT and review it in weekly or monthly business reviews.
- The brand should own the survey. Surveys should run from your help desk, under your rules, so the partner cannot choose which customers are asked.
CSAT vs other support metrics
| Metric | What it measures | When it is collected |
|---|---|---|
| CSAT | Satisfaction with one interaction | Right after the contact |
| QA score | Whether the agent followed the standard, scored by a reviewer | On a sample of contacts |
| NPS | Likelihood to recommend the brand | Periodically, not tied to one contact |
| First contact resolution | Whether the issue was solved without a repeat contact | From ticket data |
| Average handle time | How long a contact takes | From ticket or phone data |
CSAT and QA work best together. QA tells you whether the agent did the right things; CSAT tells you whether the customer felt it. A contact that scores high on QA but low on CSAT often points to a policy the customer dislikes, not an agent mistake. Our guide to essential call center metrics covers the rest of the scorecard.
What is a good CSAT score?
There is no single industry figure, because surveys, scales and customers differ. Call Centre Helper suggests treating anything between 75% and 85% as a good score and anything above 90% as excellent, and adds that the most useful benchmark is simply beating your own previous survey.
In practice, compare like with like: the same channel, the same contact reasons and the same survey method. A returns queue will score lower than a pre-sales chat queue, and that is not a sign the returns team is worse.
Why CSAT matters for your business
A high CSAT score means your support is doing what customers need. Satisfied customers are more likely to buy again, less likely to leave over a small price change, and more likely to recommend you. Customers who have a poor support experience are the ones who leave bad reviews and stop ordering.
For a growing consumer brand, CSAT is also the fastest early warning you have. A dip shows up within days of a shipping delay, a confusing policy or an undertrained new hire, long before it reaches churn or revenue numbers.

A CSAT improvement program in nine steps
These nine steps work whether the team is in-house or outsourced. In a BPO, the partner should run most of them and report on each.

1. Know your customers like you know your product
Group your customers by what they contact you about and what they expect: first-time buyers, subscribers, wholesale accounts. Write a short profile of each. A subscriber changing a delivery date wants speed; a customer whose order arrived damaged wants to feel taken seriously. Train agents on the difference.
2. Be present where your customers already are
Offer support on the channels your customers actually use, such as email, live chat, messaging and phone. Pick the two or three that carry most of your volume and staff them properly, rather than spreading a thin team across every channel.
3. Do not let automation replace human support
Chatbots and AI can answer simple questions like order status and store hours. But customers who cannot reach a person when they need one rate the whole experience lower. Automate the routine questions and keep a clear, fast route to an agent for everything else.
4. Build a self-service help center
Publish short articles for the questions you answer most: tracking an order, starting a return, changing a subscription. Organize them by topic and keep them current. Every question answered well in the help center is a contact that never waits in a queue.
5. Set clear rules for how your team talks to customers
Customers notice when one agent is helpful and the next is curt. Write standards for greeting, explaining a fix, setting expectations and closing the conversation, and confirm the problem is solved before ending. Your QA form should score exactly these behaviors.
6. Turn complaints into customer insights
Read the comments on every low score. Tag each one by cause: product, delivery, policy, agent or process. Review the tags weekly. If the same cause keeps appearing, it needs a fix outside the support team, and the support team is the one who can prove it.
7. Use visual tools to show, not just tell
Screenshots, short videos and screen sharing solve setup and troubleshooting issues faster than long written instructions. Keep a library agents can send in one click.
8. Track satisfaction after every interaction
Send the survey right after each call, chat or email, keep it to one or two questions, and include a free-text box. Tracking by agent and by contact reason shows who needs coaching and which processes need work.
9. Give your team permission to do more
Some problems do not fit a script. Let agents make small decisions on their own, such as a reshipment, a refund on a low-value order or a discount after a bad experience, within limits you set. Customers remember the agent who fixed it without an escalation.
Together these steps form a CSAT improvement program: measure every contact, find the causes, coach on them and fix what support cannot fix alone.
How Atidiv improves CSAT for consumer brands
Atidiv runs outsourced customer support for consumer brands in the US, UK and Australia across phone, chat, email, messaging and social. Dedicated QA analysts review tickets and calls against scorecards customized to your brand standards, with real-time feedback and weekly coaching sessions, and we measure success with metrics such as CSAT, NPS and Customer Effort Score.
- 98% average QA score
- 4.8 average CSAT
- 200,000+ customer experiences delivered
- Support in 11+ languages
- 30 to 80% of repeat queries automated, with QA guardrails
- We can scale your team by up to 200% within two weeks
See how our reviewers score and coach agents on our quality assurance page, or explore our full customer experience service. To talk it through, contact us.
Frequently asked questions
What is CSAT in BPO?
CSAT in BPO is the customer satisfaction score an outsourcing partner collects after each customer contact. Customers rate the interaction, usually from 1 to 5, and the score is the percentage who chose 4 or 5. It is tracked per agent, per channel and for the whole program, and is often a target in the contract.
How is CSAT calculated?
Divide the number of satisfied responses, usually 4s and 5s on a five-point scale, by the total number of responses and multiply by 100. If 340 of 400 respondents choose 4 or 5, CSAT is 85%.
Why is my CSAT score low even though my product is good?
Customers judge the whole experience, not only the product. Slow replies, long holds, being transferred, or a policy that feels unfair all pull CSAT down. Read the comments on low scores and tag the causes to see which it is.
What is a good CSAT score?
A common rule of thumb treats 75% to 85% as good and above 90% as excellent. The more useful comparison is against your own past scores for the same channel and contact type.
Can I improve CSAT without hiring more staff?
Yes. Build a help center for common questions, automate simple requests, tighten your response standards and coach agents on their low-scoring contacts. An outsourcing partner can also add coverage and a quality team without adding to your own headcount.
How do I get more customers to respond to CSAT surveys?
Send the survey right after the interaction, keep it to one or two questions, and make it answerable in one click inside the email or chat. Telling customers their answer will be read also helps.
What is the difference between CSAT and a QA score?
CSAT is the customer’s rating of the interaction. A QA score is an internal reviewer’s rating of whether the agent followed your standards. Tracking both shows whether a problem is the agent’s handling or the policy behind it.
Ayushi leads Customer Experience services at Atidiv with a strategic/operations-focused mindset. Her primary objective is to increase how well businesses deliver service and retain customers. She evaluates customers' journeys through marketing impact, performance metrics, and gaps to develop improved systems and processes. With a reputation for curiosity and structured thought processes.