First contact resolution, measured honestly
Almost every FCR number is flattering, because it counts tickets closed rather than problems solved. A definition that survives contact with reality.
First contact resolution is the most quoted metric in customer service and the most quietly flattering, because almost everyone measures it as "tickets closed on the first interaction". Closing is something you control. Resolving is not.
The three ways the number lies
- Closed is not solved
- An agent closes the ticket, the customer comes back three days later and opens a new one. Two tickets, both closed first time, one unsolved problem, and an FCR of 100%.
- Channel-switching is invisible
- They chat, get nowhere, and ring instead. Two systems, two records, neither aware of the other. The chat is "resolved" because it ended. This is the single biggest distortion, and it is a direct consequence of not being genuinely omnichannel.
- Deflection counts as resolution
- A bot answers, the customer gives up and asks a colleague instead. Nothing in your data distinguishes that from success.
A definition that survives contact with reality
Measure it from the customer's side, not the ticket's:
A contact is resolved on first contact if the same person does not contact us again about the same thing, on any channel, within seven days.
Three parts do the work. The same person — which requires one contact record, not one per channel. Any channel — which is where most of the hidden failures live. Seven days — long enough to catch the ones that come back, short enough to be actionable.
This number will be worse than your current one. That is the point of it.
What you need before you can measure it
- Identity that spans channels. Matching on phone number does most of the work. Without it you are counting tickets, not people.
- One timeline per contact, with chats and calls on it, so "contacted us again" is a query rather than an investigation.
- Some notion of topic. It does not need to be sophisticated — a coarse category on the conversation is enough to tell "same thing" from "different thing".
If you cannot do the first two, fix that before you report FCR at all. A number built on per-channel tickets is not a worse estimate of the same thing; it is a measurement of something else.
The counterweight metric
Track repeat contacts within 48 hours alongside it. It is simpler, harder to game, and moves in the opposite direction to anything that rewards ending conversations quickly.
If average handle time is falling and repeat contacts are rising, your service is getting worse while your dashboard improves. That combination is common enough to be worth a standing check — more on which call metrics earn their place.
What actually moves it
- Give the person answering the whole history. Most repeat contacts are someone explaining their situation for the second time to somebody who could have read it.
- Let the first responder finish the job. If they must escalate to change anything, you have designed repeat contacts in.
- Make handover carry context, so a handed-over conversation is still one contact from the customer's side. Handover done properly.
- Fix the top three repeat reasons. They are always a short list, and they are usually a knowledge problem rather than a people problem — the questions nobody could answer are your content roadmap.
One warning
Do not set FCR as an individual target. It is trivially gamed by discouraging people from coming back, which is the worst possible outcome dressed as the best possible number. Report it at team level, next to repeat contacts, and treat a change as a question rather than a verdict.