Blog · Board and management reporting · Customer service
A CSAT dashboard compares satisfaction scores with the responses behind them. Build a team and monthly view from survey exports, reconcile the counts, and distinguish changing service results from changing response mix.
A CSAT dashboard compares customer satisfaction scores, response counts, and response rates by team and period. Its central measure is satisfied responses divided by valid scored responses, with the scoring rule visible beside the result. Show the counts behind every percentage so a service manager can tell whether a change deserves action or further investigation.
The satisfaction score answers how responding customers rated a particular interaction. The response rate answers how much of the invited population supplied a usable score. Neither establishes what every customer thinks.
Customer experience is broader than this survey measure. The federal customer experience definitions describe perceptions and overall satisfaction with interactions across a service. A transactional CSAT question captures one defined part of that experience, so keep the question and interaction type attached to the dashboard.
For this example, a US business-services company asks customers to rate their satisfaction after a support case closes. The scale runs from 1 to 5. The company's documented rule counts ratings of 4 or 5 as satisfied. That is a reporting choice for this example, not a rule that every survey must use.
CSAT = Satisfied responses / Valid scored responses × 100
Valid-CSAT response rate = Valid scored responses / Successfully delivered invitations × 100
The second label matters. A company that counts any survey submission, including an unanswered satisfaction question, has a different numerator. Do not silently substitute that response rate for the one defined here.
Bring an invitation export and a response export. The invitation file needs invitation ID, case ID, team, channel, case-close date, delivery status, and invitation date. The response file needs invitation ID, response timestamp, satisfaction rating, and validity status.
Use invitation ID to connect the files. Retain one accepted response per invitation under a documented rule, such as the first valid submission. Exclude test records and undelivered invitations from the invitation denominator. Keep a count of exclusions so the reported population can be rebuilt.
Assign the reporting month from case-close date and apply a consistent response cutoff. Otherwise a late response can move a closed interaction into another month's trend. Retain the team responsible when the case closed instead of rewriting history when staff transfer.
The company below uses the same question, scale, delivery channels, and cutoff policy in both months. All delivered invitations and valid responses are included; there are no duplicate or invalid responses in these example rows. For the broader service review, keep this view beside the measures in customer-service KPIs.
These are September and October 2026 survey exports. The grouped rating columns preserve enough detail to rebuild every score.
| Month | Team | Delivered invitations | Valid responses | Ratings 1–3 | Rating 4 | Rating 5 |
|---|---|---|---|---|---|---|
| September | Service | 400 | 100 | 20 | 50 | 30 |
| September | Technical | 200 | 100 | 10 | 40 | 50 |
| October | Service | 400 | 200 | 30 | 100 | 70 |
| October | Technical | 200 | 50 | 2 | 18 | 30 |
Add ratings 4 and 5 to get satisfied responses. Divide by valid responses for CSAT, and divide valid responses by delivered invitations for the response rate. Percentages below are rounded to one decimal place; calculations use the original counts.
| Month | Team | Satisfied responses | CSAT | Valid-CSAT response rate |
|---|---|---|---|---|
| September | Service | 80 | 80.0% | 25.0% |
| September | Technical | 90 | 90.0% | 50.0% |
| October | Service | 170 | 85.0% | 50.0% |
| October | Technical | 48 | 96.0% | 25.0% |
| September total | Both teams | 170 | 85.0% | 33.3% |
| October total | Both teams | 218 | 87.2% | 41.7% |
The September totals are 600 delivered invitations, 200 valid responses, and 170 satisfied responses. The checks are 80 + 90 = 170, 100 + 100 = 200, and 170 / 200 = 85.0%. Response rate is 200 / 600 = 33.3%.
October has the same 600 delivered invitations, but 250 valid responses and 218 satisfied responses. Its CSAT is 218 / 250 = 87.2%; its response rate is 250 / 600 = 41.7%. The grouped ratings must also sum to valid responses on every source row. These identities catch missing rating categories before a chart hides them.
The overall score increased by 87.2 − 85.0 = 2.2 percentage points. Service improved by 5.0 points, and Technical improved by 6.0 points. Those team improvements are larger than the overall change because the respondent mix shifted toward the lower-scoring team.
Service supplied 100 / 200 = 50.0% of September responses and 200 / 250 = 80.0% of October responses. Technical's October score therefore carries less weight in the company result. The simple October average, (85.0 + 96.0) / 2 = 90.5%, would overstate the actual 87.2% score.
Put the monthly trend above a team table, with response counts and response rates directly beside each team score. Add a short action column with an owner and a review date. A separate view of first-contact resolution can help frame follow-up questions, but a survey score alone cannot identify the reason for a change.
Technical's participation fell while its score rose. Investigate delivery coverage and response composition before attributing that rise to a service change. The federal feedback dashboard handbook provides a useful transparency example: it describes publishing survey questions, response rates, and rating methods alongside feedback results. Its trust measure is distinct from this company's satisfaction question.
The survey owner maintains question wording, delivery rules, deduplication, and the reporting cutoff. Team managers review the underlying cases and decide which service actions to test. The service leader owns the target and decides whether a comparison remains valid after an operational change.
Use the questions a head of customer service asks to connect the score to a management review. Ask which customer groups changed, whether invitations reached the intended population, and which cases need a closer look. Write the proposed action beside the evidence; do not turn a favorable movement into a claim that a recent initiative caused it.
Upload your invitation and response exports or a Google Sheet; Covirage maps the columns, and its tools compute and check the satisfaction scores, response rates, and team totals. The external AI model explains the results 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 business context.
Include satisfied responses, valid scored responses, the resulting satisfaction percentage, delivered invitations, and the response rate. Add the reporting period, survey question, scoring rule, and team or channel. Counts make a high score interpretable and help reviewers spot thin or changing coverage.
No. CSAT divides satisfied responses by valid scored responses. A response rate divides the defined response count by an invitation or eligible population count. Label both denominators because a strong satisfaction score can accompany weak participation, and response-rate definitions can differ between surveys.
Add satisfied responses across the included teams, add their valid scored responses, and divide the first total by the second. Do not average team percentages unless every team has the same response count. Keep excluded responses outside both satisfaction counts and document why they were excluded.
Use a target agreed for your question, scale, customer population, and survey timing. Track comparable periods against that target before borrowing another company's benchmark. A change in the question, invitation process, or customer mix can break the comparison even if the displayed percentage looks familiar.