First Call Resolution: How to Measure It Honestly

First call resolution asks a simple question: when a customer contacts you about a problem, does it get solved the first time? It's one of the best single indicators of whether your support operation works. A high FCR means fewer repeat calls, lower costs and happier customers. A low FCR means people are calling back, getting transferred and repeating themselves.

The trouble is that FCR is surprisingly easy to measure badly.

The basic formula

FCR = issues resolved on first contact / total issues

The hard part is deciding what "resolved" and "first contact" mean. There are three common approaches, and each gives a different number.

Method 1: Agent-reported

The agent marks a disposition at the end of each call: resolved or not resolved.

Pros: easy, immediate, covers every call.

Cons: it measures what the agent believes, not what the customer experiences. Agents often mark calls resolved when they've done their part, even if the customer will need to call again when the refund doesn't arrive. If FCR is tied to their performance, the incentive to mark "resolved" is obvious.

Agent-reported FCR almost always comes out higher than other methods. Use it as a rough signal at best.

Method 2: Repeat-contact tracking

Look at your system data. If the same customer contacts you again about the same issue within a set window (commonly 3 to 7 days), the first contact wasn't resolved.

Pros: based on behavior, not opinion. Works across channels if your systems link calls, emails and chats to the same customer.

Cons: you need to decide what counts as "the same issue." Customers sometimes call back about something unrelated, and some don't bother calling back even though the problem wasn't fixed. They just leave.

To make "same issue" work, use consistent call reason codes, or compare call summaries. Transcripts with automatic summaries make this much easier, because you can match the reason for the second call to the first.

Method 3: Customer survey

Ask the customer after the interaction: "Was your issue resolved today?" Sometimes followed by "How many times did you contact us about it?"

Pros: measures the customer's view, which is what matters.

Cons: low response rates, and the people who respond aren't representative. Survey timing matters too. Ask too soon and the customer doesn't know yet whether the fix worked.

Which to use

Use repeat-contact tracking as your main FCR measure, check it with customer surveys, and treat agent-reported FCR as a disposition field rather than a KPI. If the three methods disagree sharply, the gap is itself worth investigating.

What lowers FCR

When you dig into unresolved first contacts, the same causes keep showing up:

• The agent couldn't do the thing. No system access, no authority to issue the credit, needs a supervisor.
• The call went to the wrong place. Routing sent a billing question to tech support.
• Missing information. The answer depended on a policy nobody could find.
• A promise without follow-through. "Someone will call you back" and nobody did.
• Problems that originate elsewhere. Shipping delays, billing errors, product defects. Support can't fix these on the call. They show up as repeat contacts until someone fixes the root cause.

What raises FCR

Give frontline staff more authority. Let them issue refunds or credits up to a limit without approval. It costs less than a second and third call.

Improve routing. Collect the reason for the call up front, through a voice agent or a better menu, and route by skill.

Fix the knowledge base. If agents need answers they can't find, they'll escalate or guess.

Close the loop on callbacks. If a callback is promised, it should be a task with an owner and a deadline.

Track repeat reasons and fix them at the source. Group repeat contacts by reason every month and send the top few to the teams that own them. Support can't fix a warehouse problem, but it can make sure the warehouse knows.

FCR and automation

When you add an AI voice agent or self-service, think about FCR for the whole journey:

• If the agent answers a question and the customer calls back the next day to ask a person the same thing, that's not resolution. Watch repeat-contact rates for calls handled by automation the same way you would for people.
• If the agent can't solve something, a warm transfer with a summary means the person who picks up can finish the job on that same call. That still counts as first contact resolution from the customer's point of view. We cover how to design this in the human handoff.
• Like average handle time, FCR for the human queue may drop when simple calls are automated, because what's left is harder. Report FCR by call type to avoid misreading this.

Building an FCR measurement you trust

Here's a practical approach that works for most teams without special software:

1. Standardize call reasons. Create a short list of reason codes and make sure every contact gets one, whether handled by a person or an agent.
2. Link contacts to customers. Use the phone number, account number or email to tie contacts together across channels.
3. Pick your repeat window. Seven days is a common choice.
4. Flag repeats. A contact is a repeat if the same customer contacts you again with the same reason code within the window.
5. Calculate FCR by reason code. Resolved first contacts (no repeat) divided by first contacts, for each reason.
6. Sample and verify. Each month, read a sample of flagged repeats to make sure they're really about the same issue. Adjust reason codes if they're too broad or too narrow.
7. Add a short survey after a sample of contacts to see whether customers agree.

A worked example

Say a team handles 1,000 billing calls in a month. 180 of those customers contact again about billing within seven days.

FCR for billing = (1,000 - 180) / 1,000 = 82%.

Digging into the 180 repeats, the team finds three main causes: refunds that take longer than customers were told (70), confusing line items on a new invoice format (60), and calls where the agent couldn't apply a credit without supervisor approval (50).

Each has a different fix. Set accurate expectations on refund timing. Change the invoice format or add a short explanation. Give agents authority to apply small credits. Fixing those three could lift billing FCR substantially without anyone working faster.

That's the real value of FCR: it points at the processes creating unnecessary calls.

FCR in self-service and AI channels

When customers use self-service or an AI agent, measure "resolution without further contact" the same way. A few things to watch:

• Containment isn't resolution. A call that stayed with the agent but ended with the customer calling back is a failure, even though it looked contained.
• Cross-channel repeats count. If someone talks to the agent, then emails the next day about the same thing, that's a repeat.
• Handoffs can still be resolutions. If the agent transfers with context and the person finishes the job on the same call, the customer experienced one contact.

Comparing FCR between human-handled and agent-handled calls of the same type is one of the most useful reports you can run after launching automation. If the agent's FCR is lower, read those transcripts first.

When FCR goes down and that's fine

Sometimes FCR drops for good reasons. If you launch a new product, contacts may rise and resolution dip while customers learn it. If you automate the simplest calls, human FCR may drop because the remaining calls are harder. Always look at the call mix before drawing conclusions.

Frequently asked questions

What's a good first call resolution rate?

It depends on your industry and call mix. Technical support will naturally have lower FCR than a line for order status. Track your own trend and compare by call type.

What window should I use for repeat contacts?

Many teams use 7 days. Shorter windows miss slow callbacks. Longer ones start counting unrelated issues. Pick one, document it and keep it consistent.

Does a transfer count as a failure?

From the customer's point of view, not if the issue is solved on the same call without them having to repeat everything. Many teams track "resolved on first call" and "resolved without transfer" as separate measures.

Written by the Voxvencer editorial team. We build and run AI voice agents for call centers and small businesses, and we write about what we see on real phone lines. Questions or corrections: info@voxvencer.com.