First Contact Resolution: The Metric That Actually Predicts Loyalty
If you could keep only one support metric, first contact resolution would be a defensible choice. It correlates with satisfaction more tightly than handle time, and it maps to something customers genuinely care about: did this get sorted, or do I have to come back?
It's also one of the easiest metrics to measure dishonestly — which is why plenty of organizations report strong FCR alongside customers who feel unheard.
Measuring it without fooling yourself
The common failure is defining resolution as “the agent marked the ticket resolved.” That measures agent confidence, not customer outcome. Two better definitions:
- Behavioral: no further contact from the same customer about the same issue within a fixed window — 7 days is a reasonable default.
- Attested: the customer confirms it's resolved in a follow-up.
The behavioral definition is cheaper and harder to game. It also catches the customer who gave up rather than contacting you again — which shows up as churn instead.
If a customer chats, gets nowhere, and then phones, many systems count two separate first contacts and score both. Resolution should be tracked against the customer's issue, not against each individual conversation.
The four levers
1. Answer availability
The most common cause of a second contact is that the first agent couldn't find a complete answer. This is a knowledge coverage problem before it's a training problem — and the failed-answer log tells you precisely where.
2. Authority to resolve
Agents who know the answer but can't act on it — no refund authority, no ability to amend an order — generate escalations that are entirely structural. Widening authority within defined limits often moves FCR faster than any content work.
3. Anticipating the next question
Many repeat contacts are a *different* question created by the first answer. Telling a customer their replacement ships in three days invites “how do I track it?” Answering both at once converts two contacts into one.
4. Confirming resolution explicitly
Closing with a direct “is there anything about this that's still unclear?” surfaces the gap while the customer is still there. It costs seconds and prevents whole contacts.
“Most repeat contacts aren't the same question asked twice. They're the question your first answer created and didn't address.”— Knowledge Agents
FCR for AI agents
The same metric applies, with one addition worth tracking separately: resolution without escalation. An AI agent that answers accurately and then hands off unnecessarily is inflating cost even though the answer was correct. Watch the gap between “answered correctly” and “ended the issue” — it's usually an actions gap rather than a knowledge one.
A realistic target
Benchmarks vary widely by complexity, so comparing your number to an industry average is mostly theatre. The useful target is your own trend line, segmented by issue type. A 60% FCR on complex technical issues may be excellent; the same number on password resets means something is badly broken.
Frequently asked questions
It depends heavily on issue complexity, so external benchmarks are of limited use. Track your own trend by issue type instead — the same percentage can be strong for technical troubleshooting and alarming for simple account questions.
Resolution rate counts whether an issue was eventually resolved. FCR counts whether it was resolved on the first interaction, without the customer needing to come back — which is what actually drives satisfaction.
