What is First Contact Resolution?

Every contact centre says it cares about resolution more than speed. Speed is on the wall in real time, and resolution is the harder thing to count. First Contact Resolution is the attempt to count it: the share of contacts whose underlying issue is actually dealt with, first time, with nothing left over. It is the best quality proxy the industry has, and it earns that status precisely because it is difficult to measure honestly.

Why it is the quality proxy

Most contact centre metrics describe effort: how many contacts, how fast, at what occupancy. FCR describes the point of the exercise. When an issue is resolved first time, three good things happen at once: the customer is spared a second call, the queue is spared a second contact, and the operation is spared the cost of doing the same work twice. When it is not, the unresolved issue comes back as more work with interest: a repeat contact from a now-frustrated customer, often longer and harder than the original would have been.

This is why a falling FCR is one of the few genuinely leading signals on the dashboard. It predicts a repeat-contact wave that has already been created but has not yet arrived, along with the satisfaction dip and the cost-to-serve rise that travel with it. The broader question of which metrics predict rather than describe is its own subject; FCR is the strongest single case.

Who decides what “resolved” means

The number is only as honest as its definition, and there are three common ways to build it:

• The agent decides. A resolution flag set at wrap-up. Cheap, universal and flattering, because people judge their own work kindly, and doubly so if the flag feeds their scorecard.
• The customer decides. A post-contact survey asks whether the issue was resolved. More honest in principle, but built on a small, self-selecting sample: the delighted and the furious answer surveys, the merely unresolved often do not.
• The absence of a repeat decides. No further contact from the same customer within a defined window counts as resolved. The most operational method, and the most sensitive to its own parameters.

The third method hides the biggest trap: the repeat window. Count repeats within a day and the customer who rings back on day three counts as a fresh contact, not a failure. Widen the window and FCR falls without a single conversation changing. Restrict matching to one channel and the customer who emailed, gave up and phoned is invisible. A centre can raise its reported FCR without resolving anything more, simply by narrowing the window or the channel match. The first discipline of FCR is the same as for any ratio: publish the definition next to the number, and never compare two figures without checking the definition held still.

The tension with handle time

Resolution takes time. Diagnosing the real problem rather than the presented one, fixing it fully, confirming the fix, and writing it up properly all lengthen the contact. So FCR lives in permanent tension with Average Handle Time: push AHT down hard enough and FCR follows it down, with the difference reappearing later as repeat volume.

A centre that targets speed and resolution at the same level of the organisation is asking its agents to choose, and they will choose the number that is enforced daily. Handle time is visible in real time and resolution is not, so speed wins by default and quality erodes quietly. The workable posture is hierarchical: decide the resolution rate the business needs, then treat handle time as the price of achieving it, managed at queue level as a cost lens rather than at agent level as a stopwatch.

One concrete example

Clearly illustrative, with no customer implied. A centre reports FCR improving for two consecutive quarters while complaints and churn tick upward. The two lines should not move that way together, so someone investigates. The FCR method turns out to be repeat-detection with a forty-eight hour window, matched within a single channel. Customers were ringing back on day three and four, and increasingly switching to email when calls failed, and every one of those follow-ups was counted as a new contact rather than a failed resolution. The metric was measuring the window, not the resolution. Under a four-week, cross-channel definition the trend had been falling for months, and the repeat wave the old number concealed was already staffed for by accident, as overtime.

The decision-intelligence angle

A falling FCR should trigger two decisions, not one. The first is immediate: staff for the repeat wave that is already in flight, because the failed resolutions of this week are the arriving volume of next week. The second is causal: find out what is failing, because low resolution is usually a symptom of something specific, a knowledge gap, agents without the authority to actually fix the issue, a broken process, or a product fault generating contacts no agent can resolve. Each cause belongs to a different owner, and choosing where to act is a decision worth making properly.

In a decision-intelligenceview, FCR enters as evidence with its quality visible: how it was measured, over what window, and how much that method flatters. Read against handle time and repeat volume rather than alone, it stops being a scoreboard and becomes what it always should have been: an early answer to the only question that matters, whether the work is actually getting done, delivered while there is still time to act on it. The decisions it triggers, and what they produced, are then worth recording, because a centre that can trace this quarter’s repeat wave to last quarter’s choices is the one that stops repeating them.