The Support Metrics Dashboard Every CX Leader Needs
The handful of numbers that actually tell you if support is healthy, on one screen.
Most support dashboards have the opposite of a metrics problem. They have forty numbers and no story. When everything is on the screen, nothing is, and the one chart that would have warned you about a bad week is buried three scrolls down. A good CX metrics dashboard is defined as much by what it leaves off as what it shows.
Here's the short list — six numbers that, read together, tell you whether support is healthy, where it's slipping, and what to do next. Each one earns its spot because it answers a question the others can't.
The six that earn their place
A leader's dashboard should fit on one screen and answer the questions you'd actually be asked in a leadership review. These six do that:
- Volume (by channel and intent). The denominator for everything else. A metric moving without context is noise — but next to volume, you can tell a rising reopen rate from a rising count of reopens. Break it out by channel and intent so spikes have an explanation.
- First contact resolution + reopen rate. Always together. FCR is the claim; reopen rate is the audit. A high FCR with a high reopen rate is first contact closure, not resolution, and the pairing is the only way to catch it.
- Response time (at a percentile). Per channel, reported as a percentile — "90% of chats under a minute" — not an average. The average hides the long-waiters, and the long-waiters are the ones who churn.
- CSAT (with response rate). The outcome metric, but only trustworthy alongside its response rate. A 95% CSAT on a 4% response rate is a survey of your happiest customers, not a measurement of your service.
- Deflection rate (net of reopens). The share of resolvable contacts resolved without a person and not reopened. It tells you how much of the load AI agents and self-service are truly carrying — and reopens keep it honest.
- Quality score. From your QA rubric, ideally across every conversation rather than a small sample. This is the metric that explains the others: when CSAT or FCR slips, quality is usually where the cause is hiding.
A dashboard's job isn't to show every number. It's to make the next question obvious. If a leader looks at it and doesn't know what to ask, it's decoration.
Why these and not the others
Plenty of common metrics didn't make the cut, and the omissions are deliberate:
| Metric | On the dashboard? | Why |
|---|---|---|
| First contact resolution | Yes | Resolution is the point of support |
| CSAT + response rate | Yes | The customer's verdict, kept honest |
| Quality score | Yes | Explains why outcomes move |
| Average handle time | No (drill-down only) | Effort, not outcome — gameable as a target |
| Ticket count alone | No | Volume only matters with context |
| Vanity NPS | No (separate cadence) | Relationship metric, not an ops pulse |
Average handle time is the most-requested exclusion. It belongs in capacity planning and in drill-downs, but as a top-line target it rewards rushing and punishes the reps assigned the hard cases. Keep it one click down, paired with FCR, where it informs without distorting.
How to read them together
The numbers only mean something in combination. A few patterns worth watching for:
- FCR up, reopen rate up. Tickets are being closed early. Coach from the conversations that came back, not the ones marked solved.
- Deflection up, reopen rate up. AI agents are delaying contact, not resolving it. Look at which intents are bouncing back.
- Response time good, CSAT down. You're fast but not resolving. Speed without resolution is the cheapest way to look busy while customers leave.
- CSAT high, response rate low. Survey bias. The unhappy customers aren't answering — go find them in the reopen and escalation data.
- Quality down, everything else lagging by a week. Quality is the leading indicator. When the rubric scores slip, the outcome metrics follow, so treat a quality dip as an early warning.
That last point is the case for the quality score earning a top-line spot: it moves first. The other five tell you what happened; quality tells you what's about to.
Keep it honest
A dashboard is only as good as its definitions. Three rules keep it from flattering you: report at percentiles, not averages, so the long tail can't hide; always show outcome metrics with their response or reopen rate, so a clean number can't mask a small or biased sample; and score quality across as many conversations as you can, because a 5% sample tells you about 5% of your customers' experience.
In BearScope, every conversation lands on one board, every conversation is scored — not just a sample — and AI actions leave an auditable receipt, so volume, FCR, response time, CSAT, deflection, and quality sit on one honest screen. See how the product brings these onto a single view, how trust and auditability work underneath, or book a walkthrough to map your own dashboard against your data.
Cut the dashboard to the six that answer real questions, read them in combination, and keep every definition honest. The point isn't more numbers — it's knowing, at a glance, what to ask next.
See it on your own conversations.
Bring your busiest day. We'll score every conversation in it.
Book a walkthrough →