How to measure customer service

Practice measures and outcome measures

Most lists of customer service metrics mix two different kinds of number.

Outcome measures tell you how customers responded: a satisfaction score, an effort score, whether they came back, what they wrote in public. They matter. But they arrive late, and they move for reasons your team doesn’t control, like price, weather or a supplier’s delay.

Practice measures tell you what your team did: whether every customer was invited to share their experience, whether every review was answered, whether every problem raised was closed. They arrive the same day, and they are entirely yours to move.

A service business that measures only outcomes can see that something is wrong without knowing which habit broke. Practice measures show where it leaks.

The four practice measures

The Reciprocity Method™ measures the uniformity of the practice, never the volume of the outcome. Its four measures each ask one question:

Ask Coverage and Moment Fidelity describe the invitation. Your defined moment is the earned moment, fixed at the point where your team gives its best care: an honest answer, a problem solved, a clean handoff. It is measured by the care given, not by how the job turned out, so the invitation reaches every customer you served. Zero moments? Make one, then ask. The guide How to ask customers for reviews covers the invitation itself.

Reply Rate describes what happens after a customer writes. The guide How to respond to reviews, good and bad covers the reply.

Resolution Rate describes the complaint. A problem is closed when the fix held and the customer has confirmed it, not when the ticket was marked done. The guide How to handle customer complaints covers the six steps.

Why Reply Rate is worth measuring

TRM Research’s 2026 study of Google review reply rates read 46,771 recent reviews of 986 local service businesses across five trades. Of those reviews, 58.9% received an owner reply, and 12.0% received one within seven days.

The bigger finding was how uneven the practice was. 39.7% of the businesses replied to 90% or more of their recent reviews, while 11.7% replied to none, and the trades themselves sat within 14 points of each other. Most of the spread sat between businesses in the same trade, not between trades. That is the kind of gap a practice measure is built to show.

The study measures a practice, not an outcome. It makes no claim that replying brings more reviews or better ones, and neither does this guide.

A worked week

An example, with made-up numbers, for a small service team:

The weakest habit here is Ask Coverage: 14 customers were never invited at all. That is the one to work on first, and it is the habit most likely to move next week, by asking more evenly rather than harder.

The outcome measures, and where they belong

Outcome measures are still worth reading. The common ones:

• Customer satisfaction score (CSAT): a short rating question after a visit or interaction, reported as the share of positive answers. Keep the survey separate from the review invitation. Every customer gets the invitation whatever score they gave, and the score never decides who receives it.
• Customer effort score (CES): how easy the customer found it to get what they needed.
• Net Promoter Score (NPS): how likely a customer is to recommend you, on a 0-to-10 scale.
• First response time: how long a customer waits for a first answer.
• Repeat business or retention: whether customers come back.
• Your public record: what customers choose to write about you, and the rating that follows.

Two rules keep them useful:

1. Outcome numbers travel up to leadership, never down to the person asking

   They belong in the owner’s or leadership’s review of the business. They are never handed to the person making the ask as a number to hit, and never broken out by person.

2. Read outcomes beside the practice

   A satisfaction score that drops in a week when Resolution Rate also dropped tells you where to look. A score read on its own tells you only that something changed.

The Method can and should help produce results, but no measure guarantees them. What it can do is show whether the practice behind them ran the same way for every customer.

Set it up in five steps

1. Pick one week and one place to count

   A tally sheet at the counter, or four columns in your job log.

2. Count customers served

   This is the base for Ask Coverage, so it has to be every customer, not only the ones the team remembers.

3. Mark each invitation

   Note whether it went out, and whether it went out at your defined moment, in your standard words.

4. Log every review and every problem

   Mark each review answered within your reply window or not, and each problem closed or open, with the date. A problem is closed when the fix held and the customer has confirmed it.

5. Find the weakest habit

   Work on that one next week, then count again.

Keep the counts at the level of the team or location. The goal is a practice that runs the same way whoever is on shift.

What measuring never does

A measure can quietly change the behavior it was meant to watch. These limits keep the practice honest:

• No count of reviews as a target, for the business or for anyone on the team
• No count per person, and no ranking of one person or location against another
• No supervisor attention used as leverage over the person making the ask
• No incentive, for customers or staff, tied to reviews
• No choosing who to invite by how happy they seem
• No asking for a star rating, or for a team member to be named
• No asking anyone to change a review

Google’s Maps user-generated content policy lists, among the things it does not allow, “Merchants requesting that staff solicit a certain number of reviews.” That line is quoted as written, without interpretation, and this guide isn’t legal advice.