First contact resolution and repeat contact rate are both attempts to answer the same operational question: are customers getting their issues resolved without having to come back? But they approach that question from opposite directions, they are calculated differently, and they can move in opposite directions even when nothing has actually changed in your contact center. Understanding why requires going deeper than the formulas — it requires understanding what each metric actually measures, where measurement breaks down, and what to do when the two metrics tell contradictory stories.
First Contact Resolution (FCR) is the percentage of contacts resolved in a single interaction, calculated as: (contacts resolved on first contact ÷ total contacts) × 100. It is measured at the level of the initial contact and asks whether that specific interaction was sufficient to resolve the customer's need.
Repeat Contact Rate is the percentage of total contact volume that comes from customers who have already contacted you within a defined observation window, calculated as: (contacts from customers who called within the last N days ÷ total contacts) × 100. It is measured at the level of the contact center's inbound demand and asks how much of your volume is generated by unresolved issues from prior interactions.
How FCR Is Measured
There are two fundamentally different methods for measuring FCR, and they produce different numbers — sometimes very different numbers.
Survey-based FCR
The simplest approach is to ask the customer directly. At the end of a call, the IVR plays a prompt: "Was your issue fully resolved today?" The customer answers yes or no. FCR is the percentage of yes responses among all respondents who completed the survey.
Survey-based FCR has one important advantage: it captures the customer's own perception of resolution, which is ultimately what drives satisfaction and repeat contact behavior. A customer who believes their issue was resolved — even if the underlying system problem persists — is unlikely to call back about it immediately.
The limitations are significant. Response rates to post-call surveys are typically low, meaning the sample may not represent the full contact population. Customers who were satisfied are somewhat more likely to complete the survey than those who were dissatisfied and simply hung up. Most importantly, survey-based FCR tends to overstate resolution: customers often say yes to the survey because the agent was helpful and the call was pleasant, even if the underlying issue was not fully fixed. Research from SQM Group has found that survey-based FCR typically runs 10–15 percentage points higher than operational FCR measured against actual callback behavior.
Operational FCR
Operational FCR avoids the survey entirely. Instead, it monitors whether a customer contacts the organization again within a defined window — commonly 1, 3, 7, or 30 days — about the same issue. If they do, the original contact is reclassified as unresolved; if they do not, it is counted as resolved.
This approach captures actual customer behavior rather than stated satisfaction. A customer who said yes to the survey but called back two days later is correctly classified as unresolved in an operational FCR model.
The limitation is the observation window: the number of days you define as your window directly determines your FCR rate. A 3-day window will produce a higher FCR than a 7-day window, which will produce a higher FCR than a 30-day window — for the same underlying contact data. This makes cross-organization comparisons nearly meaningless without knowing the window used.
Combining both methods
The most rigorous approach uses both: operational FCR as the primary metric (based on actual callback tracking) calibrated periodically against survey FCR to understand how customer perception aligns with behavior. When survey FCR and operational FCR are diverging — one improving while the other stagnates — there is usually a measurement or process explanation worth investigating.
For a broader treatment of how FCR fits into the full set of contact center measurements, see the call center metrics and KPIs guide.
How Repeat Contact Rate Is Measured
Repeat contact rate starts from a different position. Rather than evaluating individual contacts to determine whether they were resolved, it looks at the aggregate demand landing in your contact center and asks what fraction of that demand is driven by customers returning because a prior interaction failed them.
The basic calculation
Repeat Contact Rate = (number of contacts from customers who had a prior contact within the observation window ÷ total contacts) × 100
A 5% repeat contact rate means that 1 in every 20 contacts is from a customer who already called within your observation window. If your contact center handles 10,000 contacts per month and repeat contact rate is 5%, roughly 500 of those contacts are from customers returning because a previous interaction did not resolve their need.
Denominator choices and their consequences
The denominator choice matters considerably. Two common approaches:
- Any return contact: Count a contact as a repeat if the customer contacted you at all within the window, regardless of issue type. This is the broadest definition and captures all repeat demand, but may include contacts about entirely separate issues — a customer calling about a billing question one week and a product question the next week, counted as a repeat even though the issues are unrelated.
- Same-issue return contact: Count a contact as a repeat only if the customer returns with the same issue code or disposition category. This is more precise but depends entirely on whether agents are accurately and consistently coding call dispositions. Miscoded calls will be excluded from repeat contact rate even if they are genuine repeats.
Channel switching creates measurement gaps
Repeat contact rate is hardest to measure accurately when customers switch channels between contacts. A customer who calls, is told the issue is resolved, and then sends an email two days later about the same problem is a repeat contact — but almost no contact center's repeat contact rate calculation will catch this unless the organization has a unified customer identity across phone, email, chat, and SMS systems.
Most contact centers measure repeat contact rate within a single channel (typically voice). This means the metric systematically undercounts repeat demand from customers who switched to email or chat for their follow-up. The more channels your contact center operates, the more significant this measurement gap becomes. The contact center analytics guide examines cross-channel measurement challenges in depth.
The Observation Window Problem
Both FCR and repeat contact rate are defined by an observation window — the number of days within which a second contact is classified as a repeat or a failure. The choice of this window is one of the most consequential and least-discussed decisions in contact center measurement.
How window length changes the numbers
Consider a scenario where 20% of customers who contact you about a billing issue call back within 30 days about the same issue. Here is how FCR would appear under different window choices:
| Observation Window | % of Callbacks Captured | Reported FCR |
|---|---|---|
| 1 day | ~25% of all callbacks | ~95% |
| 3 days | ~50% of all callbacks | ~90% |
| 7 days | ~75% of all callbacks | ~85% |
| 30 days | 100% of callbacks | 80% |
The underlying contact center performance has not changed. The issue resolution rate has not changed. What changed is only the measurement window. An organization that shortens its FCR window from 7 days to 3 days will see its FCR number improve — but nothing about how well they are serving customers has actually gotten better.
The day-4 problem
If your FCR window is 7 days and a customer calls back on day 8, that original contact is classified as resolved in your FCR metric. But from the customer's perspective, they called twice about the same issue. The window creates a definitional boundary that the customer does not experience. This is why organizations with a 7-day FCR window can simultaneously have a repeat contact rate problem they cannot explain — the customers returning outside the window are invisible in FCR but very visible in their impact on handle time and agent capacity.
Window selection guidelines
There is no universally correct window. The right choice depends on the nature of your contacts:
- For transactional contacts (account balance, password reset, appointment booking), a 1–3 day window is usually sufficient. These issues either resolve immediately or do not resolve at all.
- For service and support contacts (billing disputes, product issues, technical problems), a 7-day window is common and widely cited as a practical standard.
- For complex issues that unfold over time (insurance claims, device repairs, investigations), 30 days may be more appropriate.
- Whatever window you choose, it must remain constant over time for trend analysis to be valid. Changing the window is a metric redefinition, not a performance change.
Why FCR Can Improve While Customers Still Return
One of the most disorienting situations in contact center management is watching FCR trend upward while repeat contact rate stays flat or rises. Several distinct mechanisms cause this divergence.
Agent coaching to confirm resolution
When agents are coached to ask "Is there anything else I can help you with today?" and to confirm that the customer's issue has been resolved before ending the call, survey FCR improves — customers leave the call feeling that it was resolved. But if the underlying system issue, process gap, or information failure that generated the contact still exists, customers will call back. The agent did not lie; the issue genuinely seemed resolved in the moment. The fix didn't stick, or the resolution required a back-end action that didn't complete.
Containment in self-service raises the difficulty mix
Self-service tools — IVR, web portals, chatbots — handle a category of contacts that are simple, clear, and easily resolved without agent involvement. As self-service improves, these easy contacts are resolved without ever reaching an agent. The remaining contacts that reach agents are progressively harder. Even if agent performance is constant or improving, FCR on agent-handled contacts may decline simply because the mix is harder. Conversely, if you are measuring total FCR including self-service completions, self-service improvements can mask declining FCR on agent contacts.
Short-abandon exclusions inflate FCR
Most FCR calculations exclude contacts that abandoned before reaching an agent — the caller hung up after 10 or 20 seconds, often because the wait time was too long. These contacts are excluded from FCR on the grounds that no service was attempted. But the customer still had an unresolved need. Many of them call back. Those callbacks generate volume and appear in repeat contact rate without having been counted in the original FCR denominator. FCR looks clean because abandoned contacts are excluded; repeat contact rate rises because those callers return.
Observation window narrowing looks like improvement
As described above, shortening the FCR observation window from 7 days to 3 days will immediately produce a higher FCR number. If this change happens during a period of process improvement, leaders may attribute the FCR improvement to the process work. Separating measurement changes from genuine performance changes requires documenting when and why the window definition changed.
Channel switching hides repeat contacts
A customer who calls, is told the issue is resolved, and then emails about the same issue two days later does not appear as a repeat contact in a voice-only FCR measurement. Voice FCR is clean because the callback happened in a different channel. But the customer needed a second interaction, the email agent spent time on the same issue, and the customer's experience was one of unresolved need. Repeat contact from channel-switchers is real operational load even when it is invisible in single-channel FCR.
Using FCR and Repeat Contact Rate Together
The most diagnostic approach is to track both metrics simultaneously and interpret them in combination. Four scenarios emerge from the 2×2 matrix of high/low FCR and high/low repeat contact rate.
| FCR | Repeat Contact Rate | Interpretation | Action |
|---|---|---|---|
| High | Low | Genuine resolution. Agents are resolving contacts effectively, customers are not returning. This is the healthy state. | Maintain. Monitor for trend changes as contact mix evolves. |
| High | High | Measurement mismatch. FCR definition is probably too narrow, the observation window is too short, or channel switching is hiding repeat contacts from the FCR calculation. | Audit FCR methodology. Extend the window. Add cross-channel contact matching. Investigate the most common repeat contact reasons. |
| Low | Low | Silent churn signal. Customers are not getting resolved but they are also not calling back — possibly because they gave up, switched provider, or resolved the issue elsewhere. Low CSAT will likely confirm the problem. | Check CSAT immediately. Conduct customer journey research. Look for volume drops that might indicate customers bypassing the contact center entirely. |
| Low | High | Genuine resolution problem. Agents are not resolving contacts on the first attempt, and customers are returning. The metrics are aligned and both point to the same operational failure. | Root cause analysis by issue type. Identify the top repeat-contact drivers. Investigate whether the failure is agent knowledge, policy, system access, or escalation path. |
The low FCR + low repeat contact rate combination is the scenario that most often surprises operations leaders, because it is the one where customers are suffering but the contact center appears to be performing acceptably. Repeat contact rate does not show a problem because the customers who were not resolved simply stopped contacting you. This is a customer attrition signal dressed up as operational stability.
Common Measurement Mistakes
1. Using survey FCR as the primary metric without operational validation
Survey-based FCR overestimates actual resolution because customers tell agents the issue is resolved when it is not. When survey FCR is the only FCR measurement, organizations set targets and track trends based on a number that is systematically inflated. The risk is real improvement being obscured when it is actually below the threshold needed to prevent repeat contacts. At a minimum, calibrate survey FCR against operational callback data quarterly. If the gap between survey FCR and operational FCR is widening, something has changed in how agents close calls or how customers respond to surveys.
2. Changing the observation window and calling the result an improvement
This is one of the most common accidental misrepresentations in contact center analytics. A team changes from a 7-day FCR window to a 3-day window — perhaps because the new platform calculates it that way by default — and FCR rises by several points. If this change is not documented and disclosed, leaders will believe performance improved. The correct practice is to document window changes as methodology changes, run both definitions in parallel during the transition period, and recalculate historical FCR under the new window to preserve trend validity.
3. Excluding abandoned contacts from FCR without tracking them in repeat contact rate
Abandoned calls are typically excluded from FCR because no agent interaction occurred. This is methodologically reasonable. But those customers had unresolved needs, and many of them will call back. If the repeat contact rate calculation also excludes these callers — because they had no completed contact to track against — repeat demand from abandonment-driven callbacks is invisible in both metrics. Track abandonment separately and cross-reference high-abandon periods with subsequent repeat contact rate changes to understand the relationship.
4. Measuring FCR at the contact level rather than the customer-issue level
A customer who calls three times about the same issue over 10 days generates three contacts. If your window is 7 days, the first contact and second contact will link; the first and third may not (depending on when day 7 falls relative to the second and third calls). Measuring FCR at the contact level can miss repeat contacts that arrive just outside the window from each prior contact. Measuring at the customer-issue level — flagging any customer who contacts about the same issue class more than once within 30 days — gives a cleaner picture of resolution failure regardless of how many touches it takes.
5. Ignoring repeat contact rate by call disposition
An aggregate repeat contact rate hides enormous variation by issue type. A billing dispute may have a 15% repeat contact rate while a password reset has a 1% repeat contact rate. Tracking repeat contact rate at the call type or disposition level identifies which issues are systematically failing to resolve. These are the categories where root cause analysis — is it a knowledge gap, a system limitation, a policy that prevents resolution on first contact? — will have the highest operational impact.
6. Not accounting for planned multi-contact journeys
Some contact types are inherently multi-touch. An insurance claim may require an initial notification call, a follow-up call after investigation, and a third call for resolution. None of these is a repeat contact driven by a failure to resolve — they are expected steps in a process. If these contacts are included in repeat contact rate without adjustment, the metric will be inflated by planned multi-touch interactions. The fix is to flag multi-step process types and either exclude them from repeat contact rate or analyze them separately.
FAQ
Use our free First Call Resolution (FCR) Calculator to quantify your FCR cost impact and model savings from a 5% improvement — no email required.
For the full set of contact center performance metrics and how they connect to each other, see the call center metrics and KPIs guide. For how CSAT interacts with FCR and repeat contact rate in the low-FCR, low-repeat-contact-rate scenario, see the CSAT deep dive. For the analytics infrastructure needed to measure both metrics reliably, see contact center analytics.
