First contact resolution is the only queue metric worth a target

Set a target on average handle time and it will fall, because an agent can always end a call sooner. Set a target on calls per hour and it will rise, for the same reason. Set a target on first contact resolution and the only way to move it is to actually resolve the contact. That asymmetry is the entire argument for it.

What counts as resolved

The definition has to be settled before the measurement means anything, and there are only two honest options.

• No repeat contact within a window — usually seven days, on any channel, about the same issue. Measured from the contact record rather than from anyone’s judgment.
• The agent’s own disposition — the agent marks whether it was resolved. Cheap to collect, and worth roughly what it costs, because the person marking it is the person being measured.

The first is the one to use. The second is only useful as a cross-check: where the two disagree badly, that is a coaching conversation or a broken process.

The window, and why the channel matters

A seven-day window catches most genuine repeats without absorbing unrelated contacts. The harder part is the channel: a customer who calls, gets nowhere, and then emails has not had their contact resolved, but a system that measures voice and email separately will record two first contacts and congratulate itself.

FCR is only meaningful if it is measured across channels against a customer identity. If your platform cannot join a call to an email from the same person, that is the thing to fix before setting the target.

What a low score usually means

In most centers, poor FCR is not an agent problem. The recurring causes are structural:

• The agent cannot see the system that holds the answer.They take a message, someone else looks, the customer calls back.
• The agent cannot authorize the fix.Everything above a threshold goes to a supervisor queue that runs on a different clock.
• The process genuinely has two stepsAn order that has to ship, a change that takes a day to propagate. This is not a failure, and it should be excluded from the measure rather than counted against it.
• The customer was routed wrongly.The transfer lost the context, so the second agent starts from nothing.

Three of those four are fixed by access and authority rather than by training.

The trade nobody mentions

Improving FCR usually increases handle time, and that is the correct trade. A call that runs two minutes longer and does not recur has cost less than two calls that each ended promptly. If handle time is also under a target, the two will fight, and the one with the louder report will win.

If you are going to set a target on FCR, it is worth explicitly relaxing the handle time target at the same time, and saying why.

Reading it honestly

• Segment by queue. A billing query and a technical fault do not have the same achievable ceiling.
• Exclude the genuinely multi-step processes, and keep the exclusion list short and visible so it cannot quietly grow.
• Watch the repeat-contact reasons, not just the rate. The reasons are where the next fix is.
• Do not chase 100%. A center reporting it has a definition problem, not an operation worth copying.

Where to start

Measure it before targeting it, for a full quarter, across channels, with the exclusions agreed in writing. Most centers find the number is lower than the agent-marked version by ten points or more, and that the gap is concentrated in two or three queues.

Our call center solutions hold contact history against the customer rather than against the channel, which is the part that makes a cross-channel FCR measurable at all. If you want to see what your own repeat rate looks like joined up, tell us what you are running.