Customer experience dashboard: measures and examples

A customer experience dashboard puts survey scores beside complaints and resolution times, with a separate denominator for each measure. It should show both what customers reported and what happened in service records. A company average without segment counts can hide the experience that needs attention.

Start with a decision and two exports

Consider an illustrative US office-supplies distributor reviewing July through September 2026. Its customer experience manager wants to decide which segment needs a service review before the next quarter. The dashboard combines an invitation-and-survey export with a support-ticket export.

The survey file needs an invitation ID, customer ID, segment at invitation, invitation date, response date, question version, and numeric satisfaction answer. Keep unanswered invitations so the response denominator survives. Use the same question and scale throughout the comparison: Pew Research Center explains that trend measurement requires consistent wording and attention to question context.

The ticket file needs a unique ticket ID, customer ID, segment, created timestamp, closed timestamp, status, and an approved complaint flag. Remove names and message text when those fields are unnecessary for the calculations. A customer dashboard can connect these measures to broader customer reporting; this example stays with experience and service outcomes.

Define each measure before drawing it

The distributor uses a satisfaction scale from one through five. Answers of four or five count as satisfied. Blank and invalid answers stay visible as exclusions but never become a zero score.

Satisfaction rate = Satisfied answers / Valid scored answers

Valid response rate = Valid scored answers / Eligible invitations

Complaint incidence = Tickets flagged as complaints / Tickets created

Mean resolution hours = Total elapsed hours on eligible closed tickets / Eligible closed tickets

For this dashboard, an eligible invitation is delivered to an eligible customer during the quarter. Each retained invitation has at most one accepted answer. Complaint tickets belong to the quarter's created-ticket population. Resolution time includes only those tickets created during the quarter and closed by September 30; open tickets remain in a separate count.

This last rule matters. The average describes completed work, not the expected final resolution time of every ticket. For a fuller operational view, use the customer service dashboard, which owns the service-reporting layout.

A worked dashboard by segment

The source summaries are below. Elapsed hours are calendar hours between creation and closure, summed across the eligible closed tickets.

Applying the definitions gives the dashboard measures. Percentages are shown to one decimal place where needed, and resolution times to two decimals.

Small-business satisfaction is 60 / 80 = 75.0%. Its response rate is 80 / 200 = 40.0%, while complaint incidence is 24 / 240 = 10.0%. Those measures describe different populations even though they appear on the same row.

If the source table occupies A1:H5 in Excel, put these formulas beside the first segment and fill down:

=C2/B2
=D2/C2
=F2/E2
=H2/G2

For underlying response rows, Microsoft documents COUNTIFS for counting rows meeting multiple criteria. With segment in Survey column B and validated score in column C, satisfied answers for the segment named in A2 are:

=COUNTIFS(Survey!B:B,A2,Survey!C:C,">=4",Survey!C:C,"<=5")

Microsoft's AVERAGEIFS documentation supports calculating a mean from rows meeting several conditions. If Tickets contains only the quarter's created tickets, with segment in B, status at the September 30 cutoff in C, and elapsed hours in D, use:

=AVERAGEIFS(Tickets!D:D,Tickets!B:B,A2,Tickets!C:C,"Closed")

Covirage is not affiliated with Microsoft or Google.

Prove the totals and read the differences

Company satisfaction is 162 / 200 = 81.0%, not an unweighted average of the segment percentages. Complaint incidence is 37 / 500 = 7.4%. Resolution time is 1,400 / 425 = 3.29 hours, using the unrounded calculation before display.

The segment counts must reconcile to the original exports after documented exclusions. Created tickets also reconcile as 425 closed plus 75 still open = 500. A favorable closed-ticket average deserves less confidence if the open work contains older, harder cases.

Small business is the first review candidate: it has the lowest observed satisfaction, lowest survey coverage, and highest complaint incidence. This alignment does not show that slower resolutions caused lower scores. Review the complaint categories and respondent coverage before choosing a service change. Keep account renewal and adoption questions in the neighboring customer success dashboard.

Change one input and watch the result

Suppose a later review corrects the small-business satisfied count from 60 to 68 after checking the original answers. Every other input stays unchanged. This is a numerical variation, not a claim that a service action produced improvement.

Small-business satisfaction rises by 10 percentage points, while company satisfaction rises by four percentage points. Response rate, complaint incidence, and resolution time do not move. A dashboard should make that distinction easy to see rather than implying that every aspect of experience improved.

Where it goes wrong

• One denominator everywhere. Answers, invitations, tickets, and closed tickets are different populations. Label each rate with its own base.
• Survey coverage disappears. A high score among a small or selective respondent group does not describe every customer. Show invitation and valid-answer counts beside the score.
• Customer segments change between files. Use one declared segment assignment for the comparison date so the survey and service rows refer to the same groups.
• Completed tickets hide the backlog. Keep the open count visible and review its age before interpreting a faster resolution average.
• Different questions become one trend. A wording, scale, or invitation-timing change can break comparability. Mark that break instead of connecting the figures as one continuous series.

Compare your own experience measures in Covirage

Upload your survey and ticket summaries, or bring a Google Sheet with the same reporting definitions. Covirage maps the columns; its tools compute and check the supplied measures, and the external AI model explains the differences with every figure traced to its rows. Create a free Covirage account and bring a sample of your own data.

See Covirage for customer service teams for the supporting industry context.