The five call metrics worth a dashboard

Most call dashboards measure what is easy to count rather than what changes a decision. Here are five numbers worth a place on a screen, and three popular ones that mostly are not.

The five

1. Unanswered calls, by hour

Not the daily total — the hourly split. The total tells you there is a problem; the split tells you it is lunchtime and the hour after close, which is a staffing decision or an after-hours agent, not a general exhortation to answer the phone faster.

This is the number most likely to be missing entirely, and the one most likely to be uncomfortable. Start here.

2. Never called back

Of the calls nobody answered, how many were never returned? This should be zero and almost never is. It is the clearest money-left-on-the-table number in the business, and it needs no new software to fix — only a queue with owners.

3. Median time to first contact

Median, not mean. From enquiry to a human or agent actually speaking to them. One good afternoon will hide a terrible weekend inside an average. Why this decides who gets the customer.

4. Outcome mix, by agent and by campaign

Not "calls made". What happened: reached and spoke, no answer, busy, voicemail, callback requested, wrong number, not interested, do not call.

The interesting signal is the shape. An agent with unusually many "not interested" and few "callback requested" is having a different conversation from one with the reverse. A campaign with a high "wrong number" rate has a list problem, not a script problem.

5. Repeat contacts within 48 hours

Somebody ringing back is telling you the first contact did not finish the job. It is the honest counterweight to every metric that rewards ending calls quickly, and almost nobody reports it — which is how FCR numbers stay flattering.

The three to be careful with

Average handle time

Measures speed, and rewards ending calls rather than resolving them. If you track it, track it beside repeat contacts or it will quietly make your service worse.

Calls per agent per day

Fine as a capacity signal, terrible as a performance target. It is trivially gamed by making more, worse calls — and an agent working a harder list will always look lazier.

Total call minutes

A cost line, not a performance metric. Rising minutes with rising conversion is good news; the number alone tells you nothing.

Two rules for the dashboard itself

Every metric needs an owner and a decision. If nobody can say what they would do differently at a worse number, it is decoration. Take it off.

Derived numbers must be derived. If "in the calling queue" can be hand-set by anyone, it stops describing your data and starts describing someone's opinion from March. Outcome, stage and queue state are three different things, and reports only reconcile when they are stored as three different things.

A reasonable first dashboard

1. Unanswered calls by hour, this week versus last.
2. Never-called-back count, with names attached.
3. Median time to first contact.
4. Outcome mix for the week.
5. Repeat contacts within 48 hours.

Five numbers, one screen, all of them actionable. Add the sixth only when someone can say what they would do about it.