First call resolution rate: measuring it on your customers' real callbacks

A customer who calls back about the same thing was not served on the first call. Advisor-entered status, instant survey, same number within 7 days: the usual FCR measures mislead. The method for a reliable rate, and how Once & Done in Raisetalk applies it to 100% of calls.

A queue of blue call bubbles, two of them coral and joined by an arc

What is first call resolution (FCR) and how do you measure it reliably?

  • First call resolution (FCR, or Once & Done) measures the share of requests settled without the customer having to call back about the same thing within a given period.
  • The usual measures mislead. The status entered by the advisor says what they believe, the survey says what the customer thinks in the moment, and "the same number calls back within 7 days" also counts callbacks about something else.
  • A reliable rate rests on four conditions: identify the customer, compare requests rather than phone numbers, set a window, and count a call only once its window has elapsed.
  • Once & Done, in Raisetalk, applies this method to 100% of identified calls: two labels on every call, the customer's history on the evaluation page, and a report showing where callbacks concentrate.
  • A repeat contact describes the customer's journey, not an advisor's work. The levers are in the processes: promised callbacks, requests left pending, information given to the customer.

What first call resolution really measures

A customer calls about a problem. Either it is settled and the customer does not call back, or they call back two days later about the same thing. The first call resolution rate counts the share of requests in the first case. It goes by several names: FCR (First Call Resolution, or First Contact Resolution), or Once & Done ("once, and it's finished").

Few indicators say so much at once. Every callback is one more call to handle, and therefore a workload and a cost. It is also a customer who had to make the effort of calling back, explaining again, sometimes waiting again. And it is often the trace of a process that did not hold: a promised callback that never came, a request passed to another department that did not move forward, a piece of information the customer lacked to act alone.

That is why FCR appears on every customer service dashboard, alongside average handling time and answer rate (see our article on customer relationship KPIs). Yet it often appears there with a figure that is not based on actual callbacks.

Why the usual measures mislead

Three methods dominate, and each answers a different question from the one being asked.

MethodWhat it really measuresIts limit
Status entered at the end of the call ("resolved", "pending")What the advisor believes they achievedA call declared resolved may be followed by a callback: the resolution did not hold, and the status will never know
Survey after the call ("was your request resolved?")What the customer thinks, in the moment, when they answerFew customers respond, and they respond before knowing whether the solution works
Same number within X daysThe customers who call back, whatever the reasonA callback about another request counts as a failure, and a switchboard number blows up the figures

The third method is the simplest to set up, because it needs only a number and a date. Yet it confuses two opposite situations. The customer who calls back the next day because their router still does not work has not been served. The one who calls back the next day to change their direct debit date has been: they are returning with a new request. Telling them apart requires knowing what each call was about.

That leaves quality listening on a sample: it judges how a call was handled, but a single call does not say whether the customer called back afterwards. For that, all the calls of one customer must be linked together, which no sample allows (see what analysing 100% of calls changes).

The method: four conditions for a reliable rate

1. Identify the customer

Two calls can be linked only if they carry the same identifier: phone number, customer number, case or ticket number. The best identifier is the one that follows the customer, not the line: a customer number holds up better than a phone number when a call comes from a relative's mobile.

Some identifiers do not designate a customer: the number of a switchboard, a call-forwarding line, a test line. They tie together hundreds of unrelated calls, and must be excluded from the measure.

2. Compare requests, not numbers

When a customer calls back, the real question is: about the same thing, or about something else? So the content of the calls has to be compared: the subject, the request, the references quoted (order, invoice or case number). A callback can be a repeat contact (the same issue), a new issue, or a case that cannot be decided, which is better declared as such than forced into a box.

3. Set a window

A customer who calls back six months later about the same router is not a failure of the first call. The measure is therefore taken within a repeat contact window, most often 7 days. A short window gives a result sooner but lets late callbacks slip through; a long window sees them, at the price of a later final verdict. What matters is to set it, and not to change it from one period to the next.

4. Count a call only once its window has elapsed

This is the condition most often forgotten, and it skews the rate systematically. A call can be declared not resolved as soon as the customer calls back, sometimes the next day. It can be declared resolved only at the end of its window, when you know the customer has not called back. For 7 days, failures therefore arrive, and successes do not yet.

Take an assumption: 1,000 calls a week, of which 120 will be followed by a repeat contact, a real rate of 88%. On the evening of the last day of the week, none of its calls has finished its 7-day window, but 80 callbacks have already arrived. If you count every call whose outcome is known, that week shows 0 successes for 80 failures. Mixed with three weeks whose window has elapsed, it pulls the month's rate down to 85.7% instead of 88%: more than two points lost on a process that has not changed. And the drop repeats every time the dashboard is opened.

The rule is simple: a call enters the rate only once its window has elapsed, whether or not it was followed by a callback. More recent calls stay pending, and the figure for the end of the period is presented as provisional.

One last point of method: the verdict belongs to the original call. It is counted on its date and with the advisor who took it, even if the callback was handled by someone else.

Once & Done: the method applied in Raisetalk

Raisetalk already analyses 100% of your conversations, against your own scorecards, with each finding tied to the passage of the call that supports it. Once & Done builds on these foundations to link the calls of the same customer and say, for each one, whether the request was settled without a callback.

The starting point is a customer identifier, chosen at setup: a label in your workspace, supplied at import by your connector or by the API. Shared identifiers, such as a switchboard number, are declared excluded and count nowhere.

A contact sheet for every call

At analysis, Raisetalk sums up each call in a short contact sheet: the subject ("Order delivery"), the request of the day ("Complaint about non-delivery"), the references quoted, and the state at the end of the call (declared resolved, callback promised, waiting for the company...). When the customer has already called within the window, 7 days by default, the new call's sheet is compared with those of their five most recent calls.

Two labels on every call

LabelValuesThe question asked
ContactFirst contact, Repeat contact, New issue, UndeterminedDoes this call follow up on a recent request from the customer?
Once & DonePending, Yes, NoDid the customer have to call back about the same thing after this call?

Contact looks back, Once & Done looks forward. When the analysis ends, a call is Pending. It becomes No as soon as a repeat contact arrives within the window, and Yes when the window elapses with no repeat contact. A callback about another issue changes nothing for the original call. These labels are set by Raisetalk alone: no one can change them.

Detection is in beta. Each verdict is shown with this mention and a short justification, which the manager can read back and check against the original call.

On the evaluation page, the customer's history

On the evaluation page, the Customer history card shows the verdict and its justification, the Open the original call link for a repeat contact, and the customer's other calls. The Customer summary button opens Discovery on that customer alone, over the last 12 months, with a question ready to ask: topics raised, requests, what was settled, what remains open (see querying your calls in natural language).

In the conversation list, the Repeat contacts view groups callbacks about the same thing, and the Customer field finds all of a customer's calls. The video below shows this journey, then the report.

Video: repeat contacts, the customer's history and the Once & Done report in Raisetalk

The Once & Done report: seeing where callbacks concentrate

The Once & Done report opens from the Understand page, like the other analytical reports. It answers five questions: how many requests are settled on the first call, how many calls are callbacks about the same thing, where those callbacks are, who calls back, and after how long.

Once & Done report in Raisetalk: Once & Done rate, repeat contact rate, contacts and median time to repeat, with the summary of the period
  • The Once & Done rate: Yes calls out of the calls whose window has elapsed. Contacts still pending are counted separately, with the date until which the rate can move.
  • The repeat contact rate: the share of calls that are a callback about the same thing, with the share of callbacks that Raisetalk could not classify.
  • The contacts and customers of the period, and the median time between a call and its repeat contact.
  • The trend by day, week or month, compared with the previous period. A week whose window is still running is marked provisional: its point is hollow, and a note says until when the rate can move.

A rate is shown only from 10 contacts (known contacts for the Once & Done rate, measured contacts for the repeat contact rate): below that, the report gives the number of contacts rather than a percentage that would mean nothing.

By team and by advisor: describing, not ranking

The table gives, by team or by advisor, the contacts, the Once & Done rate and the repeat contact rate. It is sorted by team and then by name, and no colour qualifies a rate. A customer also calls back because a delivery is late, because a department is waiting for a validation, because a promised callback did not come: a repeat contact describes the customer's journey. For the same reason, the report summary never receives an advisor's name or a customer identifier, and the PDF does not contain the view by advisor.

The state at the end of the call: the most direct lever

This is the most useful tile for taking action. For each of the six states recorded at the end of the call (declared resolved, waiting for the customer, waiting for the company, callback promised, escalated, not handled), it gives the Once & Done rate. A call declared resolved and then followed by a repeat contact signals a resolution that did not hold; a callback promised followed by a repeat contact, a commitment the customer waited for in vain.

Once & Done rate by state at the end of the call in the Raisetalk Once & Done report: declared resolved, waiting for the customer, waiting for the company, callback promised, escalated, not handled

The customers who call back, and when

The breakdown of customers by number of contacts shows the share of customers who call several times, and the ten most frequent customers are listed with their repeat contacts. An identifier with more than 100 contacts over the period is flagged with a Switchboard? badge: it is probably a shared number, to be excluded. Delays are read in bands, from under an hour to over 7 days: a callback within the hour can signal a dropped call or missing information, a callback several days later a request that has not moved forward.

Like the other reports, it can be compared, shared, exported as a PDF and scheduled (see the summary and PDF export of reports).

Reducing callbacks: where to start

The Once & Done rate does not improve by asking advisors to "resolve better". It improves by dealing with what makes customers call back, and the report says where to look.

  1. Keep promised callbacks. If the "callback promised" state is the one most often followed by a repeat contact, the customer calls back before being called. Note each callback with its time slot, and track in the report the share of promised callbacks followed by a repeat contact.
  2. Move forward the requests waiting for the company. A case passed to another department, with no announced timescale, almost always comes back through a new call. Giving the customer a timescale and keeping to it cuts off a good share of these callbacks.
  3. Check what "resolved" means. A call declared resolved and then followed by a repeat contact deserves a listen: was the solution the right one, did the customer know what to do next? The evaluation page links the two calls in one click.
  4. Look at callbacks within the first hour. They often signal a dropped call, a lost transfer or a missing document. These are the easiest callbacks to avoid.
  5. Cross-reference with your call reasons. The labels of your workspace can be filtered in the report: a reason that concentrates repeat contacts points to a process to review, not a team.

Quick callbacks: do not penalise the advisor twice. When a customer calls back a few minutes after a first call, introducing themselves again, re-verifying identity or rephrasing the request no longer makes sense. Raisetalk can waive the criteria you designate on the second call, within a delay set with you (60 minutes by default): the criterion is N/A, does not count in the score, and the evaluation page shows the earlier call that opens the waiver. The report reviews the criteria waived.

What the measure does not tell you

An honest measure also states its limits.

  • Only identified calls count. A customer who calls back on another channel, or from another number, is not seen. The more the identifier follows the customer rather than the line, the more complete the measure.
  • A classification can be wrong. Two close requests may be taken for the same one, two wordings of one problem for two issues. The justification of each verdict and the share of undetermined callbacks make it possible to judge.
  • The end of the period remains provisional. As long as a call's window is running, it is pending. A period of one to three months reads better than a week.
  • A shared identifier distorts everything. A switchboard that has not been excluded inflates contacts and callbacks: it is the first thing to check when a figure is surprising.

On the data side, the contact sheet follows your retention rules: if transcript purging is active, its content is erased once the window has passed, and only the state at the end of the call, the verdict and its justification remain. Customer identifiers stay on screen, reserved for users who have access to the report: neither the PDF nor scheduled sends contain them.

Key terms

  • First call resolution rate (FCR, Once & Done): share of calls whose window has elapsed and which were not followed by a callback from the customer about the same issue
  • Repeat contact: call from a customer who calls back about the same issue as an earlier call, within the window
  • Repeat contact rate: share of calls that are a repeat contact; it is read on the follow-up call, whereas the Once & Done rate is read on the original call
  • Repeat contact window: period during which a callback is attached to the original call, 7 days by default
  • Known contact: call whose window has elapsed, and whose Yes or No verdict is therefore final
  • Customer identifier: data that links the calls of the same customer, such as a phone number, a customer number or a case number
  • Quick callback: callback from the same customer within a short delay, which can waive certain criteria on the second call

Frequently asked questions

How do you calculate the first call resolution rate?

Divide the number of calls that were not followed by a callback about the same issue by the number of calls whose repeat contact window has elapsed. More recent calls stay out of the calculation until the end of their window: counting them would wrongly lower the rate, because a failure shows as soon as the customer calls back, whereas a success shows only at the end of the window.

What window should you choose to measure FCR?

Seven days is the most common window. A shorter window gives a result sooner but lets late callbacks slip through; a longer window sees them, but the verdict arrives later. In Raisetalk, it can be set from 1 to 30 days. What matters is to keep the same window from one period to the next, so that rates remain comparable.

What is the difference between the Once & Done rate and the repeat contact rate?

The Once & Done rate is read on the original call: did this customer have to call back after this call? The repeat contact rate is read on the follow-up call: is this call a callback about the same thing? The two are not exactly complementary, because they do not cover the same calls or the same dates.

Why is the rate for the last week missing or provisional?

Because the window of its calls is still running. A call enters the rate only once its window has elapsed; before that, it is pending. The Once & Done report marks these weeks as provisional and says until what date the rate can still move.

Should you track the first call resolution rate by advisor?

With caution. A customer often calls back for reasons that do not depend on the person who answered: a late delivery, a case waiting in another department, a callback promised by someone else. The rate by advisor serves to understand a journey, call by call; the levers for improvement are mostly in the processes. That is why Raisetalk presents advisors without ranking and never names them in the report summary.

Do you need particular telephony software?

No. Each call must arrive with a customer identifier, passed on by Raisetalk's connectors or by its API. The identifier is chosen at setup, and shared numbers are excluded from the measure.

Getting started

The first call resolution rate is too important an indicator to rest on a declaration or a number that calls back. Once & Done measures it on what your customers actually said, call after call, and shows where callbacks concentrate. It is the logical next step after the new Raisetalk interface: solid foundations, made more useful still. To see the feature in two minutes, watch the Once & Done video; to measure it on your own calls, try Raisetalk for free.

Blue tokens spread over a board divided into four zones, a single coral token set apart in one corner
Quality Monitoring 12 min

Evaluation disputes: what they tell you about your quality scorecard

A heavily disputed criterion is not necessarily a badly written one. Dispute rate and acceptance rate read criterion by criterion, the cause selected by the arbitrator, response time as a median: the method for knowing which criteria to review and which to explain, and how the Raisetalk Disputes report puts it into practice.

A row of blue chairs facing each other in pairs, one coral pair in the centre
Quality Monitoring 12 min

Tracking coaching: checking that debriefs take place, and what they change

Evaluating is not coaching. The number of debriefs says neither whether they were completed nor which advisors were seen. Stages, times as a median, team coverage, criteria worked on before and after: the method for tracking coaching, and how the Raisetalk Debriefs report puts it into practice.

A coral label is tied by four threads to four distinct objects set side by side
Quality Monitoring 16 min

Quality monitoring: what the term really covers, and why two tools carrying that label do not do the same job

Scoring advisors, checking regulatory compliance, listening to the voice of the customer, coaching a sales force: four disciplines circulate under the "quality monitoring" label. They do not have the same users, the same scorecards or the same standards of proof. The framework for telling them apart, why a tool does not carry over from one to another, and how to write a specification that does not mix them up.

A coral speech bubble with jagged edges, picked out in the middle of a stream of blue bubbles
Use Case 12 min

Customer incivility: identifying it in calls to alert and support advisors

An angry customer is not an uncivil customer. Insults, threats, discriminatory remarks: incivility targets the advisor as a person, and most of the time it goes unnoticed. A three-step method to identify it on 100% of calls, alert the right manager and support teams, without measuring emotion.

Three screens of the new Raisetalk interface: dashboard, evaluation page, contestation panel
Conversational Analysis 7 min

The new Raisetalk interface: conversational analysis moves up another gear

Raisetalk unveils its new interface, delivered the weekend of 1 November: more modern, clearer and, above all, more productive, built on the foundations that make it strong. A dashboard for every role, every score proven in one click, one way to search everywhere, and everything you rely on still in place.

Thousands of small blue tiles, of which only about thirty are lit up in coral
Quality Monitoring 14 min

Analysing 100% of calls: what the statistics say, and what it really costs

With 30 calls listened to a month, an 80% compliance rate in fact lies anywhere between 63% and 91%, and a ten-point drop cannot be told apart from chance. The calculation vendors do not show, the gap from which an action plan is justified, what analysing every call really changes, what it does not solve, and its cost per evaluated conversation.

Upload a call. Read its analysis. Five minutes.

Sign up in seconds, with no credit card and no cookies. A welcome tour walks you through the key principles, then you add your first conversation.

Early access · Your feedback shapes the product