Metrics guide

CSAT explained

Q: What is a good CSAT score for a contact centre?

CSAT benchmarks vary significantly by industry, channel, and survey methodology. As general reference points: a CSAT above 85% is considered strong for inbound contact centre interactions; 75–85% is typical; below 70% consistently indicates a systemic quality problem. By industry: financial services typically achieve 78–88%; retail 80–89%; utilities and telecoms 65–78% (higher complexity, higher customer frustration baseline); healthcare 82–90%. Voice interactions generally score 3–7 points higher than digital channels for the same operation, as agent empathy is more easily conveyed in voice.

Q: What is the difference between CSAT, NPS, and CES?

CSAT (Customer Satisfaction Score) measures satisfaction with a specific interaction, on a 1–5 or 1–10 scale. It answers: 'Was this interaction satisfactory?' NPS (Net Promoter Score) measures overall brand loyalty with a single question: 'How likely are you to recommend us? (0–10)'. It answers: 'Do you like us as a company?' CES (Customer Effort Score) measures how much effort the customer had to expend, typically on a 1–7 scale from 'very low effort' to 'very high effort'. It answers: 'How easy was it to get this resolved?' For contact centre operations, CES is often the strongest predictor of repeat contact rate: high effort interactions produce the most callbacks and reduce FCR.

Q: How does CSAT relate to FCR and AHT in contact centre planning?

FCR and CSAT are strongly correlated: operations with FCR above 75% typically report CSAT 10–20 points higher than operations with FCR below 60%. AHT and CSAT have a more complex relationship: very short AHT (agent rushing to close the interaction) correlates with lower CSAT; very long AHT does not consistently produce higher CSAT. The optimum is the time required to resolve the query completely on the first contact, neither rushed nor padded. For WFM planning, CSAT can be used as an output validation metric: if headcount reductions produce faster handling but lower CSAT, the efficiency gain is illusory because repeat contacts increase volume.

CSAT (Customer Satisfaction Score) measures how satisfied a customer was with a specific interaction. It is the most widely used customer outcome metric in contact centres, and one of the most frequently misread. This guide covers measurement, benchmarks, limitations, and the relationship between CSAT, FCR, AHT, and staffing.

CSAT formula

Standard CSAT calculation

CSAT = (satisfied responses ÷ total responses) × 100

What counts as "satisfied"?

On a 1–5 scale: scores of 4 and 5 (top-2-box)

On a 1–10 scale: scores of 8, 9, and 10 (top-3-box)

Use a consistent top-box definition across surveys: switching from top-2-box to top-3-box inflates CSAT by 15–25 points.

Worked example (1–5 scale)

100 survey responses received

42 scored 5 ("Very satisfied")

31 scored 4 ("Satisfied")

CSAT = (42 + 31) ÷ 100 = 73%

CSAT vs NPS vs CES

MetricQuestion askedWhat it measuresBest use

CSATHow satisfied were you with today's interaction? (1–5 or 1–10)Satisfaction with a specific interactionAgent and interaction quality; operational quality monitoring at contact centre level

NPSHow likely are you to recommend us to a friend or colleague? (0–10)Overall brand loyalty and advocacyCompany-level sentiment; relationship tracking; less sensitive to individual contact performance

CESHow much effort did you have to put in to get your issue resolved? (1–7: low to high effort)Friction and ease of resolutionPredicting repeat contacts and churn; CES is the strongest predictor of repeat call rate

CSAT collection methods

Post-call IVR

8–18% response rate

Immediately after call end, the system dials the customer and asks 1–3 questions via recorded prompts. Response rates: 8–18%. Advantage: immediate, captures mood closest to the interaction. Disadvantage: intrusive, as some customers decline or give the lowest score simply to end the call.

Best for: High-volume voice operations; operational-level real-time feedback

Post-interaction SMS or email

12–25% (SMS) / 5–15% (email) response rate

Short survey sent within minutes of call end via SMS, or within hours via email. Response rates: 12–25% for SMS; 5–15% for email. Advantage: non-intrusive; slightly higher quality responses than IVR. Disadvantage: requires mobile number or email address on record.

Best for: Omnichannel operations; operations with validated customer contact data

In-app or post-chat survey

20–40% response rate

Served immediately at the end of a chat session or in-app interaction. Response rates: 20–40%, the highest of all methods, because the customer is already using the interface. Captures digital channel CSAT with low friction.

Best for: Digital channels (chat, in-app, WhatsApp, web messaging)

CSAT benchmarks by industry

Top-2-box CSAT on a 1–5 scale for contact centre interactions. Voice CSAT is consistently higher than digital for the same operation.

IndustryVoice CSATDigital CSAT

Financial services (banking, insurance)80–88%74–83%

Regulated interactions; compliance scripting can feel impersonal, so empathy and resolution are the key CSAT drivers

Retail and e-commerce82–90%76–85%

WISMO and returns are the highest-volume contacts; fast resolution on these drives strong CSAT

Utilities (energy, water, broadband)65–76%60–72%

High frustration baseline, as customers often contact about failures. CSAT ceiling is structurally lower; effort and empathy matter most

Telecoms and ISP62–74%57–68%

Technical complexity and repair wait times suppress CSAT. First-time fix rate for field service is the primary driver

Healthcare83–91%77–85%

High stakes; empathy and clear information are primary CSAT drivers; hold time has outsized negative impact

Hospitality (hotels, travel)78–87%71–80%

Complaint handling has the highest CSAT variability; well-handled complaints can score higher than routine contacts

Public sector68–79%62–73%

Long SLAs, telephony quality issues, and complex cases structurally suppress CSAT vs. commercial operations

How CSAT connects to FCR, AHT, and staffing

FCR and CSAT

FCR is the strongest operational driver of CSAT. Operations with FCR above 75% report CSAT 10–20 percentage points higher than those with FCR below 60%. Customers who need to call back for the same issue score the second interaction significantly lower, regardless of how well the second agent handles it. Improving FCR is the highest-leverage CSAT improvement available.

AHT and CSAT

The relationship between AHT and CSAT is not linear. Very short AHT (below typical for the query type) correlates with lower CSAT, because agents are perceived as rushing to close. Very long AHT does not consistently produce higher CSAT. The optimum is the time genuinely required to resolve the query. For WFM planning, artificially capping AHT to reduce headcount cost typically produces a CSAT drop and a repeat contact volume increase, undermining the efficiency saving.

Wait time and CSAT

ASA (average speed of answer) has a moderate but consistent negative effect on CSAT. For inbound voice, every 30 seconds of additional wait time reduces CSAT by approximately 1–3 percentage points. For chat, customers are more tolerant of initial queue time but react strongly to agent response latency during conversation. Service level investment directly protects CSAT.

CSAT as a staffing validation metric

CSAT can reveal whether headcount reductions are saving cost or simply shifting it. If a reduction in agent headcount produces faster handling (lower AHT) but lower CSAT and higher repeat contact rate, the saving is illusory, because repeat contacts increase volume and cost. Monitor CSAT alongside AHT when evaluating the operational impact of staffing changes.

Common CSAT measurement pitfalls

Measuring only contacted customers

Customers who abandoned the queue are not surveyed. Abandonment typically correlates with dissatisfaction. A CSAT of 85% from answered contacts may reflect a true CSAT of 68% if 20% of customers never reached an agent.

Small sample sizes producing false precision

A CSAT of 83% from 12 responses has a confidence interval of ±25 points. A CSAT of 83% from 1,200 responses has a confidence interval of ±2 points. Report sample size alongside every CSAT figure.

Coaching agents to manipulate scores

Agents who are instructed to 'ask for a 5' or tell customers 'anything less than a 5 affects my bonus' produce CSAT data that is useless for operational improvement. Score-soliciting is also a GDPR and ICO concern in some implementations.

Switching scale or top-box definition mid-trend

Moving from a 1–5 scale to a 1–10 scale, or from top-2-box to top-3-box, changes the CSAT score by 15–25 points. Never compare across the change point without restating the series on the new basis.

CSAT questions

What is CSAT in a contact centre?

CSAT (Customer Satisfaction Score) measures how satisfied customers were with a specific interaction, collected via a post-interaction survey. It is calculated as the percentage of respondents who gave the top 1–2 ratings on a 1–5 scale (or top 3 on a 1–10 scale). CSAT measures reaction to a specific event — different from NPS (overall brand loyalty) or CES (perceived effort).

What is a good CSAT score for a contact centre?

Strong: above 85% on a 1–5 top-2-box basis. Typical: 75–85%. Warning signal: below 70% sustained. By industry: retail and healthcare 80–90%; financial services 78–88%; utilities and telecoms 62–76%. Voice interactions typically score 3–7 points higher than digital for the same operation.

What is the difference between CSAT, NPS, and CES?

CSAT: satisfaction with a specific interaction. NPS: overall brand loyalty and likelihood to recommend. CES: how much effort the customer had to expend. For contact centre operations, CES is the strongest predictor of repeat contact rate, because high-effort interactions generate the most callbacks and are the clearest signal of FCR failure.

How does CSAT relate to FCR and AHT in contact centre planning?

FCR is the strongest CSAT driver: above-75% FCR operations typically report CSAT 10–20 points higher than below-60% FCR operations. AHT and CSAT are non-linear. Rushing calls to hit an AHT cap reduces CSAT, while excessive AHT does not consistently improve it. The optimum is the time required to genuinely resolve the query on the first contact.