Customer Experience Metrics: 12 KPIs, Formulas and How to Use Them

The three groups of customer experience metrics

Perception metrics come from asking customers a question. They are fast to collect and easy to explain, but they only reflect the people who answered, at the moment they answered.

Operational metrics come from your own systems: the help desk, the phone platform, the product logs. Nobody has to fill in a survey, so the data covers every customer, but it tells you what happened, not how the customer felt about it.

Outcome metrics are what the business finally sees: customers who stay, leave, buy more or recommend you. They move slowly and have many causes, which is why you need the other two groups to explain them.

Perception metrics: CSAT, NPS and CES

Customer satisfaction score (CSAT) asks how satisfied the customer was with one interaction, usually on a 1 to 5 scale. CSAT = number of 4 and 5 answers ÷ total answers × 100. If 180 of 220 respondents choose 4 or 5, CSAT is 82 percent. Send it right after a ticket closes or a milestone is reached, while the customer still remembers the details.

Net Promoter Score (NPS) asks how likely the customer is to recommend you, from 0 to 10. NPS = percentage of promoters (9–10) − percentage of detractors (0–6). With 50 percent promoters and 20 percent detractors, NPS is +30. It measures the relationship as a whole, so run it on a fixed schedule, not after every ticket.

Customer Effort Score (CES) asks how easy it was to get something done, often on a 1 to 7 scale. CES = average score, or the share of answers at 5 and above. Of the three, it is the easiest to act on, because a high-effort answer points to a specific step in a specific workflow.

Do not add the three scores together or swap one for another from quarter to quarter. Each answers a different question, and a trend only means something if the question, the scale and the audience stay the same.

• CSAT: one interaction, sent within an hour of resolution.
• NPS: the whole relationship, quarterly or twice a year, at account level in B2B.
• CES: one task or workflow, sent after onboarding steps, setup flows or self-service attempts.

Operational metrics: speed and resolution

First response time (FRT) is the time between a customer's first message and the first human reply. Report the median and the 90th percentile together. The median shows the normal experience; the 90th percentile shows the customers who waited long enough to become angry.

Average resolution time is the time from opening to final resolution. Measure it in business hours if your team does not work around the clock, and exclude time spent waiting on the customer, otherwise the number punishes the team for the customer's delay.

First contact resolution (FCR) = tickets solved in the first interaction ÷ all tickets × 100. If 640 of 1,000 tickets never need a second reply, FCR is 64 percent. FCR usually tracks satisfaction more closely than speed does: customers forgive a slower answer that ends the problem more readily than a quick one that doesn't.

Repeat contact rate is the share of customers who come back about the same issue within seven days. It catches the tickets that were closed but not solved, which FCR can miss when agents close tickets too early.

Outcome metrics: churn, retention and customer lifetime value

Customer churn rate = customers lost in the period ÷ customers at the start × 100. Customer retention rate is the reverse: (customers at the end − new customers) ÷ customers at the start × 100. Starting with 400 customers, adding 60 and ending with 430 gives a retention rate of 92.5 percent.

Customer lifetime value (CLV) = average revenue per customer per year × gross margin × average customer lifetime in years. A customer paying $6,000 a year at 80 percent margin who stays four years is worth $19,200 in gross profit. CLV turns experience work into a number finance teams recognise.

In subscription businesses, add net revenue retention. It shows whether customers who stay also buy more. A good experience tends to show up there first, well before it shows in logo churn.

How to connect experience scores to revenue

The test of a CX metric is whether it predicts an outcome. Group customers by their score, for example detractors, passives and promoters, then follow each group's renewal rate and expansion over the next two or three quarters.

If promoters renew at 95 percent and detractors at 70 percent, you have a case for investment, and a way to estimate what moving 10 percent of detractors is worth. If there is no difference between the groups, the score is measuring something your customers do not act on, and you should change the question or the moment you ask it.

Present the result as a relationship, not as proof. Plenty of other things change during a renewal cycle, and a finance team will trust the next analysis more if the first one did not overclaim.

Which customer experience metrics to track first

A small team should start with four: CSAT after support interactions, first contact resolution, customer churn rate and one relationship score (NPS for B2B accounts, CES if most customers use self-service). That gives one number from each group plus a quality check on support.

Add metrics only when someone owns them and a decision depends on them. A dashboard of twenty CX metrics nobody acts on costs reporting time every week and changes nothing.

• Support-led teams: CSAT, FCR, first response time, repeat contact rate.
• Product-led teams: CES on key flows, activation rate, churn, NPS.
• B2B account teams: account-level NPS, time to first value, gross and net revenue retention.

Common mistakes when measuring customer experience

Changing the survey audience and calling the new score an improvement. A score that jumps because you stopped surveying free users has not improved.

Reporting averages only. One bad week or a few very long tickets can hide inside an average; medians and percentiles show them.

Setting targets that agents can game. An FCR target with no repeat contact check rewards closing tickets, not solving problems.

Collecting comments and never reading them. Theme the free-text answers every month and send each theme to an owner; the comments are where the reasons are.