First call resolution measurement: the data you need

Short answer

You cannot tell from one call whether the problem was solved on the first contact. FCR (First Call Resolution) is measured by combining three things: what the contact was about, the same customer's later contacts, and an agreed observation window. The agent saying "that's solved", or the call ending on a friendly note, replaces none of them.

So FCR is a definition problem first and a reporting problem second. An FCR figure with no written definition cannot be compared across teams and measures something different every quarter.

"Solved" versus actually resolved

At the end of a call the agent closes the contact as "solved". That is their own judgement. Actual resolution shows up as the customer not coming back about the same issue. There is usually a gap between the two, for different reasons:

• The agent answered, but the customer's real question was something else
• The promised action (refund, technician visit, correction) was never carried out
• The customer did not understand the answer and came back through another channel, such as chat
• The issue was outside the agent's authority, but no transfer was made either
• The status was set to "solved" so the open-case count would look smaller

The last reason matters most: tie FCR to the agent's own flag and it quickly becomes a managed number.

The data FCR needs

Computing FCR takes the fields below. None of them comes reliably from a call transcript — they come from telephony, the CRM or the ticketing system.

1. 1Customer identifierPhone number or customer ID. If one customer uses several numbers, you need a rule that joins them.
2. 2Contact reasonThe category or contact type the agent selects. Without it you cannot tell "same issue" from "different issue".
3. 3Contact time and channelCall, chat, email — if the customer comes back through another channel, that is a repeat contact too.
4. 4Closure statusThe agent's flag. It is not FCR itself; keep it as a separate field for comparison.
5. 5Status of the promised actionRefund, order change, technician visit — was it actually executed in the system?

Writing the definition: five decisions

1. 1WindowWithin how many days does a repeat contact count? A short window inflates FCR; a very long one assigns new problems to old contacts. The period should fit the contact type.
2. 2Same-issue ruleDo only repeat contacts about the same reason count, or any repeat contact? The second is simpler, but counts a customer who returns with a new question as "unresolved".
3. 3ChannelsIs a contact that starts by phone and continues in chat included?
4. 4ExclusionsContacts where the customer asked for a callback, or where a second step is planned (a call after a technician visit), are flagged separately.
5. 5SourceIs FCR computed from system data, a customer survey or both? Each has weaknesses; the source belongs in the report.

Illustrative example: one week, three FCRs

The figures below are an illustrative example, not real company data. In one week there were 200 first contacts. Agents closed 188 of them as "solved". Within the next 7 days, 38 customers got in touch again: 26 about the same issue, 12 about something else.

Same week, same calls, three different numbers. Asking which one is "right" misses the point; what is right is writing the chosen definition down and keeping it stable. The 7-point gap between 94% and 87% is valuable in its own right: the 14 contacts the agents called solved but the customer came back about are the best review list you have.

What the call itself adds

A transcript does not prove FCR, but it can explain why FCR is low. When calls that led to a repeat contact are compared with the first call, these patterns usually appear:

• The need was not clarified — the agent answered the first question they heard
• The next step was vague: "we'll check", with no when and no who
• The agent did not check that the customer understood the answer
• The customer's second question was not recorded
• A transfer was made, but the information did not reach the next team

That list is also a set of criteria you can write into your call standard. Checking them on every call takes the problem the FCR report shows down to the level of behaviour.

How to show FCR in a report

• Print the definition next to FCR: window, same-issue rule, channels, exclusions
• Show the agent-flag figure separately, for comparison
• Split by contact type — simple enquiries and technical faults have different FCR
• Look at agent-level FCR only with enough contacts; a percentage from a handful of contacts is noise
• When the definition changes, do not join the old and new periods on one chart

A weekly review routine

An FCR figure is only useful when the cases behind it are reviewed. For a small team, a weekly rhythm like this is enough:

1. 1Monday: the listPull last week's contacts that were flagged "solved" but came back about the same issue within the window.
2. 2SelectionPick 10–15 cases across different agents and contact types; do not take one agent's calls only.
3. 3ReviewListen to or read the first call in each case. Record the cause in one category: need not understood, wrong answer, action not carried out, failed transfer, process gap.
4. 4OutcomeAgent-related causes go to individual coaching; process causes go to the team that owns the process. If the same cause repeats two weeks running, it is a system problem.

The review takes an hour or two, but it turns the FCR report from a "the number went up or down" discussion into a list of concrete fixes.

Limits

• One call's transcript does not prove the issue was solved — later contacts have to be checked
• A transcript does not confirm that a CRM action was taken or a refund went through
• No repeat contact does not always mean resolution: the customer may simply have given up and moved to a competitor
• Survey-based FCR reflects only the customers who answered
• FCR is not a metric the agent owns alone; process and systems affect it too

What Vexvon Audio Analyzer offers

Vexvon Audio Analyzer does not compute FCR: the customer's later contacts and the case status live in your telephony and CRM. Its role is to give evidence for the "why" above.

• You write steps such as "the need was clarified", "the next step was specific" and "the agent checked the answer was understood" into your call standard
• On every call each step is judged as met, partial, missed or not applicable, with a comment and the transcript lines behind it
• A short summary of each call saves time for whoever works through the repeat-contact list

Uploading the calls that led to repeat contacts and checking them against these steps surfaces the behaviour behind a low FCR. More: Vexvon Audio Analyzer. The balance between call length and quality is covered in AHT vs call quality.

First step

Write your FCR definition on one page: window, same-issue rule, channels and exclusions. Then compute last week's FCR by the agent flag and by system data side by side, and review 10 calls from the gap. This article belongs to the support quality and coaching section. To try it on your own recordings, get in touch.

Further reading on this topic: call transfer quality, call recording analytics integration.