What call center attrition benchmarks are
Call center attrition benchmarks are published annualized turnover rates for contact center floors, and the honest starting point is that far fewer of them exist than the industry talks like they do. Three national series carry real weight: ContactBabel's US and UK Contact Centre Decision-Makers' Guides, and the CCAP Attrition and Retention Survey in the Philippines. There is no published series for the Caribbean, Latin America or India, and there is no credible global average at all.
This tool scores your floor against those published series where one exists. Where one does not, it says so instead of handing you a default that reads like a benchmark. For the Caribbean and Latin America it shows our own working range, labeled at the point of use as a Call Force Global estimate rather than a published measurement, because Call Force Global holds no proprietary attrition dataset and will not present one as though it did.
Correction, August 2, 2026. This tool previously published a set of attrition bands by geography, sourced on the page to a research firm, an industry association and Call Force Global operational data. None of those three attributions held. The bands have been replaced with the published figures where a published figure exists, and removed where none does. The attrition cost index withdrawal notice sets out what was withdrawn and why.
Published contact center attrition by market
| Market | Published annual attrition | Source and edition |
|---|---|---|
| United States | 31% mean, 24% median (year-end 2023); a third of respondents above 30% | ContactBabel, US Contact Center Decision-Makers' Guide, 2024 edition, 189 US contact center managers |
| United Kingdom | 15% to 32% across 2003 to 2023, recently 23% to 24% | ContactBabel, UK Contact Centre Decision-Makers' Guide, 2024 edition |
| Philippines | 45% total in 2022 (31% voluntary, 14.1% involuntary); 43% on 2023 data reported May 2025 | CCAP Attrition and Retention Survey, conducted by Willis Towers Watson across 145 member organizations |
| Philippines, wider IT-BPM | 30% to 40% | IBPAP president and CEO Jack Madrid, June 2023. A press remark, not a published study. |
| Caribbean, Latin America, India | No published figure | No survey series exists. Any band you see, including one with our name on it, is a vendor estimate. |
Two figures still in circulation are worth naming so you can stop repeating them. A 60 to 70 percent Philippine rate is CCAP's own description of the era before 2016 and is not current. A 28 percent Philippine figure from the first half of 2023 covers six months and must never be annualized. The published Philippine trend has been downward, not upward. If you want the longer treatment of what the data does and does not support by geography, the call center attrition guide carries it, and the Caribbean attrition delta sets out the structural case for a nearshore gap that we cannot evidence with a number.
Why the vertical adjustment is our estimate, not a benchmark
The vertical multipliers in this tool encode an ordering we see in our own operations: quota-driven outbound and collections are the hardest seats to hold, and regulated verticals with a heavier training investment are the stickiest. That ordering is a Call Force Global estimate. No published attrition series is broken out by the verticals this tool lists, so there is nothing to source it to.
Because of that, the multiplier is applied only to our own working ranges. It is never applied to a published figure. Multiplying a ContactBabel mean or a CCAP total by a number we made up would produce something that looks published and is not, which is exactly the failure this tool is being corrected for. If you operate across several verticals, weight your own measured rates by seat count rather than reaching for a multiplier at all.
The hidden cost of attrition
Direct replacement cost is the obvious bucket: recruiting, training time, equipment. Only one component of it has a published anchor. SHRM puts cost per hire at a $4,683 average against a $1,244 median (2022 Talent Access Report, n=472) and its 2025 Recruiting Benchmarking report puts the nonexecutive median at $1,200, down from that $1,244. The roughly fourfold gap between mean and median is right skew, so quote whichever one you are prepared to defend and say which it is. Training and equipment costs have no published figure we know of. Use your own.
The bigger bucket is ramp inefficiency. A new hire takes time to reach full productivity, and produces less than a tenured agent while getting there. This tool prices that at six weeks at roughly 70 percent lost productivity, which is a Call Force Global working assumption rather than a published measurement. Your own ramp curve will differ, sometimes between two floors in the same city, so measure it. To run the whole model on your own inputs, use the turnover cost calculator.
Knowledge loss is the third bucket, and it has no defensible dollar figure at all. Senior agents who leave take with them the edge cases, the escalation routes, and which clients tolerate what. It shows up in CSAT and first-contact resolution two quarters after a churn wave rather than on a P and L. The full cost of attrition breakdown covers it qualitatively, which is the only honest way to cover it.
How to lower attrition
Call Force Global treats attrition as a tech-enabled problem before a wage problem. Three levers, in the order we work them. We are not attaching a percentage improvement to any of them, because we do not hold the dataset that would let us prove one.
- AI voice screening at hire. Every candidate is scored on a composite of English fluency, communication and grammar, and anyone under the threshold is filtered before a human interview. The point is that likely wash-outs never enter the production pipeline, which is cheaper than losing them at week six. See the AI QA call center breakdown.
- Daily AI QA coaching. Traditional QA samples a small share of calls, which means most calls are never reviewed and a quiet failure mode is invisible by definition. AI QA scores every call every day, so agents get a feedback loop instead of a monthly surprise.
- Performance dashboards visible to agents. When agents can see their own trajectory on CSAT, handle time and SLA contribution, they self-correct before a supervisor has to intervene, and the resignation conversation stops being a surprise on both sides.
On top of those, the structural argument for nearshore fronter perimeter design is that the highest-churn work sits where wage elasticity absorbs it, while the work you most need continuity in stays where continuity is easiest to buy. That is an argument about design, not a claim about a measured gap between markets.
FAQ
What is the average call center attrition rate in 2026?
There is no credible global average, and only a few markets have a published series at all. US contact center attrition runs a 31 percent mean and a 24 percent median on year-end 2023 data, with a third of respondents above 30 percent (ContactBabel, US Contact Center Decision-Makers' Guide, 2024 edition, 189 US contact center managers). The UK has run 15 to 32 percent across 2003 to 2023, recently 23 to 24 percent (ContactBabel, UK Contact Centre Decision-Makers' Guide). Philippine contact center attrition was 45 percent in total in 2022, made up of 31 percent voluntary and 14.1 percent involuntary, easing to 43 percent on 2023 data reported by CCAP in May 2025 (CCAP Attrition and Retention Survey, conducted by Willis Towers Watson across 145 member organizations). No published series exists for the Caribbean, Latin America or India.
Why might nearshore call center attrition be lower than offshore?
We cannot put a sourced number on that gap, because no survey series covers the Caribbean or Latin America, and we are not going to invent one. The structural case is this: same-timezone daytime shifts avoid the 11 to 13 hour inversion that drives graveyard-shift fatigue, employer pools per metro are smaller so there is less poaching within commuting distance, and contact center work is more often treated as a career than a stepping stone. Those are reasons to expect a difference. They are not a measurement of one.
How do I calculate the true cost of call center attrition?
Multiply your annual attrition rate by your seat count to get expected departures, then multiply by your cost per departure. Cost per departure is your direct replacement spend plus ramp drag, the productivity you pay for but do not receive while a new hire climbs to speed. This tool prices ramp drag as loaded hourly rate times 6 weeks times 40 hours times 0.7. A 25-seat floor at 35 percent attrition and $18 per loaded hour loses roughly $26,460 per year to ramp drag alone, before recruiting spend and supervisor coaching time. The 6 weeks at roughly 70 percent lost productivity is a Call Force Global working assumption, not a published measurement, so replace it with your own measured ramp curve.
Which call center vertical has the highest attrition?
No published attrition series breaks the market out by the verticals this tool lists, so there is no sourced answer to give. In our own operations the hardest seats are quota-driven outbound and collections, and the stickiest are the regulated verticals where the training investment is heavier. The vertical multipliers in this tool encode that ordering, they are Call Force Global estimates rather than published measurements, and they are applied only to our own working ranges and never to a published figure.
Can technology lower call center attrition?
We believe so, and we are careful about how we say it, because we do not hold an attrition dataset that would let us prove a percentage. AI voice screening at hire filters candidates likely to wash out before they enter the production pipeline. AI QA that scores every call rather than a 2 percent sample turns coaching into a daily loop instead of a monthly surprise. Dashboards that show agents their own trajectory take the surprise out of the conversation before it becomes a resignation. Those are mechanisms we run. Any percentage improvement attached to them, from us or from anyone else, is an estimate.
What attrition rate should I target for a nearshore team?
There is no published nearshore series to set a target against, so treat any number as a working goal rather than a benchmark. The more useful move is to drop the single blended figure and target attrition by tenure band instead. If most of your loss lands inside the first 90 days you have a hiring and onboarding problem, which is the cheaper one to fix. If it lands around month nine it is a wage and progression problem, which is not. Judge yourself against your own trend, because attrition definitions vary enough that cross-company comparison is often meaningless.
Keep going
Compare your operational KPIs to industry: the call center KPI benchmark dashboard scores AHT, FCR, ASA, CSAT, occupancy, and SLA against the same 2026 vertical bands. See what nearshore agents actually cost with the Caribbean wage index across Jamaica, Trinidad, Belize, and Colombia. Or model your funnel economics with the cost-per-qualified-meeting calculator. You can also run your headcount through the nearshore cost calculator to see what a 10-seat team would cost on Caribbean rates.
For deeper reading: true cost of attrition, Caribbean attrition delta 2026, AI QA in call centers, nearshore fronter perimeter, or our customer support outsourcing service. Pricing assumptions sit on how pricing works. To request a quote, head to contact.