First Contact Resolution: Definition, Formula, and What a Good FCR Rate Is
First contact resolution (FCR) is the share of customer issues resolved on the first interaction, with no follow-up needed.
If a customer emails once, gets a complete answer, and never has to come back about it, that contact counted toward FCR. If they have to reply, call again, or open a second ticket about the same issue, it did not. It is one of the few support metrics that customers and finance both care about, because a repeat contact costs you twice and annoys the customer once.
It is also one of the most commonly overstated numbers in support, for reasons covered below.
What Does FCR Mean?#
FCR stands for first contact resolution. You will also see first call resolution, which is the same idea confined to phone support — the older term, from an era when support was mostly a call center.
"Contact" is the superset. It covers a phone call, a live chat, an email, a web form, a social message, or a conversation with an AI agent. Teams running more than one channel should measure contact resolution rather than call resolution, or the number only describes part of the operation.
FCR in a service desk means the same thing applied to internal IT: the share of employee requests — password resets, access requests, software installs — resolved on first contact without escalating to a second tier. IT teams often call the equivalent metric first-line resolution or first-level resolution, and it is measured against tier-2 escalation rather than customer callback.
How to Calculate First Contact Resolution#
The formula:
FCR = (Issues resolved on first contact ÷ Total issues) × 100
A worked example. In a week your team handles 1,400 contacts, of which 1,050 involve no follow-up on the same issue:
FCR = (1,050 ÷ 1,400) × 100 = 75%
Simple arithmetic. The difficulty is entirely in defining the two inputs, and that is where most reported FCR numbers come apart.
What Is a Good FCR Rate?#
The most cited benchmarks come from SQM Group, which has measured contact center performance for over 25 years:
| FCR rate | Assessment |
|---|---|
| 80%+ | World-class — achieved by about 5% of centers |
| 70–79% | Good |
| ~70% | The average call center |
| Below 60% | Needs attention |
SQM also reports that each 1% improvement in FCR corresponds to roughly a 1% improvement in customer satisfaction, and estimates about $286,000 in annual operational savings per 1% of FCR for an average midsize call center.
Two caveats before you benchmark yourself against those figures.
They are call center numbers. Phone support is synchronous, so an agent can keep working until the issue closes. Email and asynchronous chat naturally produce lower first-contact rates because a clarifying question ends the interaction. Comparing an email-first team to a 70% phone benchmark is comparing different things.
Industry and complexity move the number more than skill does. A utility answering billing questions and a B2B platform debugging integrations should not expect similar rates. Your own trend line is worth more than any external benchmark.
The Traps That Inflate FCR#
Most reported FCR is higher than actual FCR. Four reasons, in rough order of how much damage they do.
No follow-up window. If you count a contact as resolved the moment the ticket closes, every issue resolves on first contact by definition. FCR only means something measured against a window — commonly 7 days, sometimes 24 or 72 hours. The customer who comes back on day 8 has to count.
Self-reported resolution. An agent marking their own ticket resolved is the least reliable available signal. Verified resolution — the customer confirms, or no repeat contact occurs inside the window — is the number worth acting on. SQM's own methodology relies on post-contact customer surveys rather than agent disposition, which is precisely why their benchmark is lower than most internal dashboards.
Reopened tickets counted as new. If a customer replies about the same issue and your system opens a fresh ticket, the original still looks resolved on first contact and your volume goes up. This one silently improves two metrics while the customer experience gets worse.
Channel switching invisible. A customer who tries chat, gives up, and phones you has had two contacts about one issue. If those live in separate systems, both look like first contacts. This is the strongest practical argument for a single queue across channels.
FCR vs Ticket Deflection#
These two metrics are often quoted together and they pull in opposite directions, which almost nobody says out loud.
Ticket deflection counts contacts that never reached an agent. FCR counts contacts that reached an agent and finished immediately.
Deflect your easy questions successfully and your FCR will fall — because the simple issues that used to resolve on first contact are now handled by self-service, leaving agents with a harder residue. That is a good outcome that looks like a regression on the dashboard.
The failure mode is measuring them separately and rewarding both. A team pushed to raise deflection and FCR at once has an obvious move available: deflect aggressively, and count the customers who gave up as deflected rather than as repeat contacts. Both numbers improve, and the business gets worse.
Read them together, alongside repeat-contact rate and CSAT. If deflection climbs while FCR falls and CSAT holds steady, that is working as intended.
How AI Changes FCR — Honestly#
The optimistic version says an AI agent resolves routine questions instantly, so FCR rises. That is true only under a condition most vendors skip.
An AI that answers from a general language model will produce a confident reply to a question your content never covered. The customer accepts it, acts on it, discovers it was wrong, and comes back. That is a first contact that appeared resolved and generated a repeat contact — the exact pattern FCR exists to catch, now happening faster and at higher volume.
Grounding is what changes the outcome. Retrieval-augmented generation retrieves the relevant passages from your own content first and answers from those, citing the source. The model is reading rather than recalling, so it has nothing to invent. And when nothing matches the question, retrieval comes back empty, which is the signal to escalate rather than guess.
The handoff matters as much for FCR as the answering does. If a customer asks for a person and has to re-explain the problem, that is a second contact from their point of view even if your system logs one ticket. Denser escalates on three explicit conditions — the visitor asks for a person, the AI has failed repeatedly, or the AI detects frustration — and the conversation becomes a ticket with the full transcript attached, so an agent picks up mid-conversation rather than starting over.
Worth being precise about scope: Denser measures and improves resolution on chat, email, help page, and ticket channels. Phone is supported as a ticket channel with configurable auto-reply, not as a voice agent handling live calls — so the classic call-center FCR benchmarks above are a reference point, not a like-for-like comparison.
How to Improve FCR#
In rough order of return.
- Fix the top ten repeat questions. Pull the contacts that generated a follow-up and find the pattern. Usually a handful of topics account for most repeats, and they are almost always a content problem rather than an agent problem.
- Give agents the answer, not the search box. Most first-contact failures are an agent unable to find or verify something. Cited, retrievable answers fix more FCR than coaching does.
- Kill the re-explanation. Full context on handoff — between AI and human, between agents, between channels.
- Widen agent authority. A refund an agent cannot approve is a guaranteed second contact. Every approval threshold is an FCR ceiling.
- Measure against a window, with customer confirmation. You cannot improve a number you are currently measuring generously.
- Fix the contradictions. When two help articles disagree, agents and AI both answer inconsistently and the customer returns. Denser's answer quality insights flag contradictions across sources for exactly this reason.
Frequently Asked Questions#
What does FCR mean?#
FCR stands for first contact resolution — the percentage of customer issues fully resolved during the first interaction, with no follow-up required on the same issue. When the term is used specifically about phone support it is often called first call resolution.
What is a good first contact resolution rate?#
By SQM Group's widely used benchmark, 70–79% is good and 80% or above is world-class, which only about 5% of contact centers achieve. Roughly 70% is the average. Those figures come from call centers, so email- and chat-heavy teams typically measure lower for structural reasons rather than performance ones. Your own trend against a consistent definition matters more than the benchmark.
How do I calculate first contact resolution?#
Divide the number of issues resolved on the first contact by the total number of issues, then multiply by 100. The result is only meaningful if "resolved" is measured against a follow-up window — commonly 7 days — and ideally confirmed by the customer rather than marked by the agent who handled it.
What does FCR mean in a service desk?#
In an internal IT service desk, FCR is the share of employee requests resolved at first contact without escalation to a second-line team. It is often called first-line or first-level resolution. The concept is identical to customer FCR; the escalation path is to tier 2 rather than a customer callback.
What is the difference between FCR and first call resolution?#
First call resolution measures phone contacts only. First contact resolution covers every channel — phone, chat, email, web forms, social, and AI conversations. The terms are frequently used interchangeably, which causes confusion when a multichannel team compares itself to a phone-only benchmark.
Does a high FCR always mean good service?#
No. FCR can be inflated by closing tickets early, counting reopened issues as new contacts, or measuring resolution without a follow-up window. It can also fall for good reasons — successfully deflecting simple questions leaves agents with harder ones. Read FCR alongside repeat-contact rate and CSAT rather than on its own.
Where to Take This Next#
FCR measures what happens when a contact reaches your team. Ticket deflection measures what never reaches them, and the two need to be read together to mean anything.
Our guide to helpdesk automation covers the layer underneath both: which support work to automate, where the line should sit, and the escalation rules that stop an AI resolving things it should not.
Denser is built for the whole loop: Chatbots that retrieve from your own content and cite every answer, and a Helpdesk that catches what they should not finish, with full transcript attached so nobody re-explains anything. Start free or see answer quality.