Knowledge Center
CX Metrics Library
Definitions, formulas, benchmarks, and measurement guidance for 11 core customer experience KPIs — written for practitioners who are accountable for moving them.
Definitions, formulas, benchmarks, and measurement guidance for 11 core customer experience KPIs — written for practitioners who are accountable for moving them.
CSAT measures customer satisfaction with a specific interaction, transaction, or overall experience, captured through a direct survey question. It is the most widely used operational quality metric in customer-facing organisations.
A post-interaction survey asking customers to rate their satisfaction, typically on a 5-point scale (Very Satisfied to Very Dissatisfied). The CSAT score is calculated as the percentage of respondents selecting the two highest ratings.
CSAT provides direct, timely feedback on individual interactions — making it the most actionable quality metric for frontline coaching and process improvement. Unlike NPS, which measures relationship sentiment, CSAT measures transactional performance.
Survey timing bias — customers surveyed immediately after resolution report higher CSAT than those surveyed 24 hours later. Low response rates can skew results significantly. Aggregating CSAT across all contact types obscures performance differences between them.
NPS measures customer loyalty and likelihood to recommend — a relationship-level metric that reflects cumulative experience rather than a single interaction. It is widely used at board level and in investor reporting.
A single survey question: "On a scale of 0–10, how likely are you to recommend us to a colleague or friend?" Respondents are classified as Promoters (9–10), Passives (7–8), or Detractors (0–6). NPS is the Promoter percentage minus the Detractor percentage.
NPS correlates strongly with customer retention and growth in many sectors. It provides a simple, boardroom-legible indicator of customer relationship health. Its simplicity makes it highly comparable across time periods and against competitors.
NPS is a lagging indicator — it reflects past experience and does not predict future behaviour with precision. A neutral score (Passive) is excluded from the calculation but represents a significant portion of most customer bases. Scores vary significantly by survey method and question timing.
CES measures how much effort a customer had to invest to get their issue resolved. Research consistently shows that reducing customer effort is more strongly correlated with loyalty than increasing customer delight.
Typically a 7-point scale asking customers how easy the experience was. A high CES score (low effort) correlates with reduced churn risk and higher repeat purchase likelihood. It is particularly useful for contact centre and self-service channel assessment.
Customers who describe an interaction as requiring high effort are significantly more likely to churn and significantly less likely to recommend. CES isolates the friction dimension of an experience — distinct from overall satisfaction or loyalty intent.
CES is primarily useful for transactional contacts, not relationship-level measurement. It does not capture the emotional dimension of an experience. A low-effort, low-quality experience will score well on CES but poorly on CSAT.
FCR measures the percentage of customer contacts resolved without the need for a follow-up contact, a callback, or a transfer. It is one of the highest-impact operational metrics in contact centre management — directly linked to both CSAT and cost.
A contact is counted as resolved on first contact when the customer's issue is fully addressed in a single interaction with no requirement for the customer to contact again about the same issue within a defined window (typically 5–7 days).
FCR improvement has a compound effect: it increases CSAT (customers whose issues are resolved first time report higher satisfaction), reduces repeat contact volume (reducing cost per contact), and improves agent utilisation. The COPC benchmark is ≥70%.
FCR can be gamed by agents who mark contacts as resolved prematurely. Measurement methodology matters: customer-confirmed FCR (asking the customer) is more accurate than system-based FCR. Different contact types have very different FCR ceilings — aggregate FCR obscures this.
AHT measures the average total time spent per contact, broken into three components: Talk Time (active conversation), Hold Time (customer on hold), and After-Call Work (ACW — wrap-up time following the contact). COPC evaluates all three components independently.
AHT is the primary efficiency metric in contact centre operations. It determines staffing requirements through workforce management calculations and directly affects both cost per contact and capacity planning.
Managing AHT allows operations to handle more contacts with the same headcount — or to understand where agent time is being spent unnecessarily. Reducing Hold Time and ACW, in particular, is often achievable without affecting quality. Talk Time reduction is more complex.
AHT targets that are too aggressive drive agents to rush contacts, reduce FCR, and increase repeat contact volume. AHT should always be tracked alongside FCR and CSAT. A reduction in AHT that causes FCR to drop is not a net efficiency gain.
ASA measures how long customers wait before their contact is answered by an agent. It is a secondary measure to Service Level — where Service Level measures the percentage of contacts answered within a threshold, ASA measures the average wait across all contacts.
ASA is most useful for identifying the impact of queuing on contacts that fall outside the Service Level threshold. A low Service Level combined with a high ASA indicates significant delays for contacts not answered quickly. ASA is heavily influenced by contact arrival patterns and staffing levels at interval level.
High ASA directly correlates with increased customer frustration before the conversation even begins, making every agent's job harder and degrading CSAT independent of the interaction quality itself.
ASA is a mean average — it masks the experience of customers in the long tail of the queue. Track 90th percentile wait time alongside ASA to understand the worst-case customer experience. Abandoned contacts are excluded from ASA calculation, which understates the wait time problem when abandonment is high.
Occupancy measures the percentage of an agent's logged-in time spent actively handling contacts or in associated work. It is the primary measure of how productively agent capacity is being utilised.
Occupancy is distinct from utilisation. Utilisation is the percentage of scheduled time during which agents are logged in and available. Occupancy measures what agents are doing while they are logged in. High occupancy means agents move from contact to contact with minimal gap. Low occupancy means agents are available but idle.
Occupancy directly affects agent experience and quality. COPC defines the target occupancy range as 85–90%. Above 90%, agents have insufficient recovery time between contacts — quality declines and attrition accelerates. Below 75%, the operation is paying for capacity it is not using.
Occupancy above 92% is operationally unsustainable in any high-complexity contact environment. It appears efficient on paper but creates quality degradation and attrition costs that exceed the apparent efficiency gain. Always track CSAT and QA scores alongside occupancy.
Shrinkage is the percentage of scheduled paid time during which agents are not available to handle contacts. It is a critical workforce management input — accurate shrinkage assumptions are essential for building schedules that meet Service Level targets.
Shrinkage is divided into planned shrinkage (scheduled breaks, training, coaching, team meetings, off-phone tasks) and unplanned shrinkage (absence, late arrival, system downtime, extended breaks beyond schedule). COPC defines acceptable total shrinkage at 25–35%.
Operations that underestimate shrinkage build schedules that assume more agent availability than actually exists. The result is chronic understaffing at interval level, elevated ASA and queue times, increased occupancy beyond healthy thresholds, and declining CSAT. Tracking planned vs. unplanned shrinkage separately is essential because each requires different interventions.
Using a single annual shrinkage figure for WFM planning ignores seasonal and operational variation. Unplanned shrinkage that consistently runs above 15% signals a structural attendance or management problem — not an individual performance issue — and requires a workforce management investigation, not disciplinary action.
Quality Score is the output of a structured quality monitoring evaluation — the assessment of a specific contact against defined criteria, scored using a calibrated form by a trained QA evaluator.
A Quality Score reflects how closely an agent's handling of a contact aligned with defined standards across key dimensions: typically including greeting, needs identification, resolution accuracy, empathy, compliance, and close. The weight assigned to each dimension varies by contact type and organisation.
Quality monitoring is the mechanism through which an organisation assures that the experience customers receive aligns with the experience it intends to deliver. Without calibrated QA, quality is personality-dependent rather than system-dependent — highly variable and difficult to improve systematically.
QA scores that are not calibrated — where different evaluators score the same contact differently — are unreliable and unfair. QA programmes that use scores punitively rather than developmentally damage agent trust and produce defensive behaviours rather than genuine improvement. Track QA score trends by agent and by team lead, not just overall averages.
Conversion Rate in a CX context measures the percentage of inbound contacts that result in a commercial outcome — a sale, an upsell acceptance, a renewal commitment, or another defined commercial action. It is the metric that connects CX performance to revenue.
Traditionally a sales metric, Conversion Rate is increasingly applied to CX operations where agents are trained in consultative approaches and expected to identify and act on revenue opportunities within service interactions. It creates accountability for the commercial contribution of the CX function.
Contact centres that track and manage Conversion Rate alongside CSAT begin to position themselves as revenue contributors rather than cost centres. This reframing has significant implications for CX investment decisions, headcount justification, and the design of agent incentive structures.
Conversion Rate targets that create pressure on agents to sell during service contacts degrade CSAT if poorly implemented. The training, scripting, and performance management design must align consultative selling with service quality — not treat them as competing objectives.
Attrition measures the rate at which agents leave a contact centre — voluntary and involuntary — within a defined period. It is a leading indicator of cost pressure, training burden, and performance risk.
Agent attrition should be tracked by tenure cohort — new agents (under 6 months), developing agents (6–18 months), and experienced agents (18+ months) — as each cohort has different drivers and different cost implications. Overall attrition figures obscure these important distinctions.
High attrition creates a cost compounding effect: recruitment cost, training time, performance lag during the new-agent curve, and the institutional knowledge loss of experienced agents. Operations running at 40%+ annual attrition are typically paying more in replacement cost than it would take to address the root causes of attrition.
Tracking only overall attrition masks whether the problem is in early tenure (onboarding and culture) or late tenure (career development and management quality). Exit interview data is informative but biased — many leavers will not share their real reason for leaving in a formal exit process.