First contact resolution sounds like a simple percentage. In practice, teams can get different answers from the same queue because they count different units, use different follow-up windows, or stop the measurement at different points in the customer journey.
This evidence review separates those choices. The companion first contact resolution explainer covers the operating levers teams commonly use to improve the metric. This page focuses on what a reported benchmark actually measures, whether two figures can be compared, and how to set a transparent internal baseline.
First contact resolution benchmarks: what the data measures
First contact resolution, usually shortened to FCR, measures whether a customer's question or issue is resolved during the first interaction. APQC's phone measure defines the unit as a question or issue received by phone that is answered, addressed, and resolved by the end of the initial contact, with no future contact required for that issue. Other sources extend the term to email, chat, text, tickets, self-service, or a connected customer journey.
That difference is the first benchmark warning. A phone-only FCR figure is not automatically comparable with a cross-channel figure. A case marked resolved when an agent closes a ticket is not automatically comparable with a customer-survey answer about whether the issue was actually solved. A value without its measurement rules is a label, not a complete benchmark.
What this review can and cannot compare
The sources reviewed here provide definitions, measurement guidance, and selected benchmark observations. They do not form one harmonized industry dataset. The figures below come from different populations and collection methods, so the comparison is descriptive rather than a league table.
| Published figure | What it describes | Why it is not a universal target |
|---|---|---|
| 70% | SQM Group's aggregated 2025 FCR average across industries | A post-call survey benchmark with its own sample and survey method |
| 62% | Consumers in XM Institute's 2025 trend study who reported first-time resolution | A consumer-reported experience measure, not a system-derived case rate |
| 83% | The highest country result reported in the XM Institute summary, for India | A country comparison within that study, not a general service standard |
| 25% | The lowest country result reported in the XM Institute summary, for South Korea | A country result whose context and sample should be read in the original study |
| 70% to 79% | SQM Group's stated range for a good FCR rate | SQM's benchmark guidance, not a rule that applies to every channel or case mix |
Source note: SQM Group and XM Institute use different designs. The percentages should not be averaged or ranked as if they came from one sample.
Why FCR definitions produce different results
Most FCR disagreements come from the unit being counted, the test for resolution, and the time allowed for a return contact. Teams should document all three before they compare performance.
Contact, case, incident, or customer journey
One phone call can be counted as one contact. A ticket can be counted as one case even if the customer sends several messages inside it. An issue can span a call, a chat, and an email. A customer journey measure can connect those touches to ask whether the original problem was solved without another attempt.
APQC's definition is explicitly phone-based. Microsoft describes FCR as the share of cases resolved on the first interaction without a return contact within seven days. Gartner's summary warns that many organizations measure only one assisted-service channel even though customers experience the full journey. These are legitimate measures, but they answer different questions.
| Measurement unit | The question being answered | Typical blind spot |
|---|---|---|
| Contact | Was this interaction resolved before it ended? | Can overstate performance when the customer returns through another channel |
| Case or ticket | Was this case closed without another recorded interaction? | Depends on reopen rules and how one issue is split into cases |
| Issue or incident | Was the customer's underlying problem resolved? | Requires reliable issue linking across contacts |
| Customer journey | Did the customer need another attempt anywhere in the journey? | Needs cross-channel identity and event stitching |
| Survey response | Did the customer say the issue was resolved on the first interaction? | Survey response bias and incomplete survey coverage |
For a small support team, the practical choice is not to pretend that one definition fits every purpose. Report an operational rate for the system you can observe, and add a customer-confirmed measure when the team needs to know whether closure matched the customer's experience.
Gross FCR, net FCR, and the denominator problem
The basic formula is straightforward:
FCR = first-contact resolutions / total eligible contacts x 100
The difficult word is eligible. A gross rate can use all incoming contacts in the denominator. A net rate can remove contacts that the operation could not reasonably resolve on first contact, such as requests that require a specialist, a physical repair, or a later customer action. Medallia documents net FCR as a separate calculation that subtracts contacts not resolvable by a level-one specialist. MetricNet's sample benchmark likewise describes net FCR as applying to potentially resolvable live contacts.
Net FCR can help a manager assess the work that a queue is designed to handle. It can also make a team look stronger if the exclusions are broad or inconsistent. For that reason, publish the gross rate, the net rate if used, the exclusion rules, and the count removed from the denominator.
| Denominator choice | Include | Use with caution when |
|---|---|---|
| All incoming contacts | Every contact in the reporting population | The queue receives work that no first-line agent could resolve |
| Eligible contacts | Contacts judged resolvable by the measured team | Eligibility is assigned after seeing the outcome |
| Survey respondents | Customers who answered the FCR question | Response rates or respondent mix change over time |
| Resolvable contacts by issue family | Contacts inside a defined issue scope | Issue classification is incomplete or changes during the period |
The safest benchmark label names the numerator and denominator in plain language. "FCR: 74%" is incomplete. "74% of 4,280 eligible phone contacts were resolved without a return contact within seven days" is testable.
Recontact windows change the benchmark
The phrase "first contact" does not define how long a team should wait for evidence that the issue returned. A customer may reopen a ticket later, send a new email, or call about the same problem after an initial answer. A short window may miss those failures. A long window may attach a new, unrelated problem to the earlier contact.
Microsoft's agent metric reference uses a seven-day return-contact window. Kustomer describes a typical 24-to-72-hour tracking window for recontact, while Medallia's documentation focuses on whether the first interaction resolved the query and offers a separate net-FCR treatment. These differences are why a benchmark needs its window alongside the percentage.
| Window or rule | What it can reveal | Risk to interpretation |
|---|---|---|
| End of interaction | Immediate completion for live channels | Misses a later return about the same issue |
| 24 to 72 hours | Near-term recontact for routine questions | May not fit billing, delivery, or account issues with longer cycles |
| 7 days | A broader return-contact check, as in Microsoft's reference | Can connect a later but unrelated contact if issue identity is weak |
| Ticket reopen rule | Whether the same case returned | Fails when customers create a new case instead of replying to the old one |
| Customer survey answer | Whether the customer says the issue was solved | Coverage depends on response behavior and question wording |
Choose the window from the work. A delivery question may need a different follow-up rule from a password reset. If a single company uses multiple windows, keep them as separate named measures rather than folding them into one blended rate.
Channel scope is part of the benchmark
First call resolution is usually a voice measure. First contact resolution can include phone, chat, email, text, and tickets. The broader name does not solve the comparability problem because each channel has a different interaction shape.
Phone contact can end with a spoken answer or a transfer. Chat can include several exchanges in one session. Email may require an agent to investigate before sending a response, and a customer may not reply even when the answer was incomplete. A ticketing system may count a closure that the customer never accepted. A combined number can hide these differences.
Report channel slices before reporting a blended total. If channels must be combined for planning, disclose the channel mix and keep the same issue and recontact rules wherever the systems allow it.
| Channel or scope | Define before measuring | Useful companion measure |
|---|---|---|
| Phone | Transfer treatment, callback treatment, and repeat-call window | Customer-confirmed FCR or repeat-call rate |
| Chat | Session boundary, handoff treatment, and post-chat recontact | Chat reopen or repeat-contact rate |
| Whether one reply counts as resolution and how silence is treated | Reopen, reply, or follow-up rate | |
| Ticket | Closure, reopen, and new-ticket linkage rules | Issue-level resolution rate |
| Cross-channel journey | Identity matching and issue linking across systems | Channel-specific FCR slices |
Gartner's guidance is useful here: a channel dashboard can be operationally convenient while still missing the customer's earlier or later attempt elsewhere. The channel rate is not wrong. It is simply narrower than the journey question.
What published benchmarks actually show
Published FCR benchmarks are useful for orientation, not for copying a target into a scorecard. The source list includes several kinds of evidence:
- SQM Group reports a 70% aggregated average across industries in its 2025 research using external voice-of-the-customer measurement. It describes 70% to 79% as a good range and 80% or higher as world-class in its guidance, while also noting that industry, call type, complexity, and measurement method affect the result.
- XM Institute reports that 62% of consumers said their issue was resolved the first time they contacted customer service in its Global Contact Center Trends 2025 summary. The same summary reports country-level results from 83% in India to 25% in South Korea.
- Talkdesk describes 70% to 75% as a general benchmark in a 2024 explainer, but also says that a good rate varies by industry, issue complexity, and company goals.
- MetricNet uses benchmark distributions and percentiles rather than treating the average as the only useful comparison. Its materials distinguish FCR from first-level resolution, which can include a later callback handled by the same support level.
Those observations can coexist because they measure different things. The responsible conclusion is not that one publisher is right and the others are wrong. The conclusion is that the source method belongs in the benchmark label.
A benchmark reading checklist
Before using a published FCR number in a plan, ask:
| Check | Question |
|---|---|
| Population | Who was measured, and in which geography, industry, channel, or queue? |
| Unit | Is the denominator a contact, case, issue, customer journey, or survey response? |
| Resolution | Does resolution mean agent disposition, no reopen, no recontact, or customer confirmation? |
| Window | How long can a return contact disqualify the first interaction? |
| Exclusions | Which contacts are removed, and were they removed before or after the outcome? |
| Method | Is the figure system-derived, agent-entered, quality-reviewed, or customer-reported? |
| Distribution | Is the published figure an average, median, range, percentile, or selected example? |
| Date | When was the data collected, and when was the page or report published? |
If a source does not answer enough of these questions, use it as background rather than as a target.
Build a defensible internal FCR benchmark
An internal benchmark becomes useful when another person can reproduce it. Start with a written metric contract before pulling data.
1. Define the unit and eligible population
Choose whether the measure follows contacts, cases, issues, or customer journeys. Then define what belongs in the denominator. Write the exclusions before the reporting period begins. Do not remove a difficult case only because the outcome was poor.
2. Define the resolution test
Pick the evidence that proves resolution: a reviewed disposition, no reopen, no repeat contact, a customer survey answer, or a combination. If the operation uses more than one test, name the measures separately. A closed ticket rate is not the same as customer-confirmed FCR.
3. Set the recontact window
Use the cycle time of the issue family and document the choice. Keep a stable window while establishing the baseline. If the window changes, treat the result as a measurement break and annotate the trend.
4. Segment before averaging
At minimum, inspect FCR by channel and issue family. Add queue, language, product, customer segment, and time period where the sample supports it. A blended result can improve because easy contacts increased, even while a complex queue deteriorated.
5. Pair FCR with quality and customer feedback
FCR can rise when agents close cases early or when customers stop replying. Pair it with reopen or recontact rate, quality review, customer satisfaction, and customer effort. The customer service quality metrics guide provides the companion measurement context for this broader view.
6. Publish counts with the percentage
A rate without a denominator hides sample size. Report the numerator, denominator, excluded count, channel mix, window, and collection method. For survey results, show the response count and question wording where possible.
7. Use benchmarks as a starting point, not a verdict
The goal is a stable measure that helps a team learn. An external benchmark can frame the question. It cannot decide whether a 70% result is healthy for a particular product, queue, risk level, or customer promise.
Key first contact resolution statistics 2026
| Statistic | Figure | Source and measurement context |
|---|---|---|
| SQM Group 2025 aggregated FCR average | 70% | External voice-of-the-customer benchmark across industries |
| SQM Group stated good FCR range | 70% to 79% | SQM benchmark guidance, subject to method and case mix |
| SQM Group stated world-class threshold | 80% or higher | SQM benchmark guidance, not a universal target |
| XM Institute consumers reporting first-time resolution | 62% | Global Contact Center Trends 2025 consumer-reported result |
| XM Institute India result | 83% | Country result in the 2025 trend summary |
| XM Institute South Korea result | 25% | Country result in the 2025 trend summary |
| Microsoft return-contact window | 7 days | FCR definition in its agent metrics reference |
| SQM typical one-contact versus FCR gap | 10% or lower | SQM distinction between one-contact resolution and FCR |
| Talkdesk general FCR benchmark | 70% to 75% | Talkdesk explainer, with an explicit caveat about variation |
Interpretation note: these are reported figures from different sources. They should not be combined into one average, used as a universal service promise, or compared without their method labels.
Sources
- APQC, First contact resolution rate for phone contacts, accessed August 2, 2026.
- Gartner, How to Measure and Interpret FCR, published February 27, 2024, accessed August 2, 2026.
- SQM Group, First Call Resolution comprehensive guide, accessed August 2, 2026.
- SQM Group, First Call Resolution white paper, accessed August 2, 2026.
- XM Institute, Global Contact Center Trends 2025, accessed August 2, 2026.
- Microsoft Learn, agent metrics reference, accessed August 2, 2026.
- Medallia Documentation, How is FCR calculated?, updated March 31, 2026, accessed August 2, 2026.
- Medallia Documentation, first call/contact resolution glossary, accessed August 2, 2026.
- Qualtrics, First Contact Resolution: Definition, Measurement and Best Practices, published July 10, 2023, accessed August 2, 2026.
- Intercom, First Contact Resolution, published May 26, 2022, accessed August 2, 2026.
- Kustomer, First Contact Resolution glossary, accessed August 2, 2026.
- HDI, First Contact Resolution Calculation Guidance, accessed August 2, 2026.
- HDI, KPI Bulletin 5: First Contact Resolution, accessed August 2, 2026.
- MetricNet, First Level Resolution vs. First Contact Resolution, accessed August 2, 2026.
- MetricNet, Contact Center Metrics Essentials: How to Benchmark, accessed August 2, 2026.
- Talkdesk, First Contact Resolution, published November 13, 2024, accessed August 2, 2026.
- MetricNet, sample insourced service desk benchmark, accessed August 2, 2026.
- AWS, Amazon Connect metric primitive definitions, accessed August 2, 2026.
Related reading: Customer service metrics dashboard, customer satisfaction metrics, and first response time versus average handle time.
Frequently Asked Questions
What is a good first contact resolution benchmark?
There is no universal target. SQM Group reports a 70% aggregated 2025 average across industries and describes 70% to 79% as a good range in its guidance. That figure comes from a particular external voice-of-the-customer method. Use it as context, then compare your own result by channel, issue family, and customer journey.
Why do FCR benchmarks disagree?
They may count contacts, cases, issues, or customer journeys. They may use agent dispositions, ticket status, repeat-contact data, or customer surveys. They may also use different channels, exclusions, and recontact windows. A benchmark is not comparable until those choices are visible.
What is the difference between FCR and first-level resolution?
FCR asks whether the issue was resolved during the first interaction. First-level resolution can include a later callback or follow-up handled by the same support level. MetricNet gives this distinction directly. The measures can be useful together, but they should not share one label.
Should a team use gross or net FCR?
Use gross FCR to show the result across the stated population. Add net FCR when the team needs to assess only contacts that its defined scope could reasonably resolve. Always publish eligibility and exclusions, because net FCR can change materially when the denominator changes.
How long should the FCR recontact window be?
Choose a window that fits the issue cycle and keep it stable during the baseline. Microsoft uses seven days in one FCR metric reference. Other guidance discusses shorter windows. The important reporting practice is to name the window and avoid comparing figures that use different windows.
Can phone and chat FCR be combined?
They can be combined for a stated planning purpose, but the channel slices should remain visible. Phone, chat, email, and ticket measures have different session boundaries and follow-up behavior. A blended value can hide a weak channel or a change in channel mix.
What should accompany an FCR dashboard?
Show the numerator, denominator, exclusions, recontact window, channel mix, and data collection method. Pair FCR with recontact or reopen rate, quality review, customer satisfaction, and customer effort so that early closure does not look like successful resolution by itself.
A practical next step
Write the metric contract before you set a target: unit, eligible population, resolution test, recontact window, channels, exclusions, and source systems. If you want a second set of eyes on a support measurement or staffing plan, contact CustomerCareStaff to discuss the operating assumptions.