Caribbean Attrition Delta: Why Nearshore Voice Holds Tenure in 2026

Every senior ops person who has run an outbound voice program for more than two years has lived through the same cycle. The Philippines voice floor looked cheap on paper. The first 30 days went well. By month four, the seat that was supposed to be the same agent dialing the same prospect list was a different agent. By month nine, the floor had cycled through half its headcount. Training cost compounded, conversion tanked through ramp, and the labor-arbitrage savings the procurement deck promised had quietly evaporated into a calibration tax that never showed up on the per-hour invoice.

This is the lurking cost on every voice BPO contract. The Caribbean attrition delta is the term for the gap between that experience and what stable Caribbean nearshore voice rooms actually deliver. One side of that gap is published and one side is not, and this piece is careful about which is which. It is the reason the apparent 6 dollar per hour Caribbean premium over far-offshore voice (a Call Force Global estimate from rate cards we have seen, not a published benchmark) is often offset on a total cost of ownership basis. The point of the piece is to hand you the method, so you can run the delta on your own program rather than borrow anyone else's band.

Why attrition is the lurking cost on every BPO contract

Attrition is the most under-modeled cost component in outbound voice. Procurement teams model it as a percentage when they model it at all, and the percentage rarely shows up in the vendor pitch. The hourly rate dominates the conversation, so the procurement deck reads as labor arbitrage, and the calibration tax stays invisible until the second quarter of the contract.

Senior ops people know this. They have lived through the voice floor that cycles half its chairs a year. They have watched a 20-seat program lose 10 trained agents in 12 months and then watched the supervisor team spend the next quarter rebuilding calibration from scratch. They have seen conversion rates dip on every fresh-agent cohort. They have absorbed the recruit-screen-train-ramp cost three or four times for the same chair.

What they have rarely seen is the same chair held by the same agent for 18 months. That is the actual headline benefit of a low-attrition voice floor, and it is the buyer-felt pain that the Caribbean attrition delta is built to quantify.

What the published attrition data actually covers

Published attrition data is uneven by geography, so compare carefully. Two markets have a credible published series. The Caribbean does not, and saying so is more useful to you than a confident number nobody can check.

• United States: a 27 percent mean and a 21 percent median. A third of respondents were above 30 percent. Source: ContactBabel, US Contact Center Decision-Makers' Guide, 2024 edition, 189 US contact center managers. Note what it is not: the guide is US-only and segmented by vertical and company size, not by geography, so it says nothing about offshore or nearshore attrition.
• United Kingdom: 15 to 32 percent across 2003 to 2023, recently 23 to 24 percent. Source: ContactBabel, UK Contact Center Decision-Makers' Guide.
• Philippines: 45 percent total in 2022, easing to 43 percent on 2023 data. The 2022 figure breaks down as 31 percent voluntary and 14.1 percent involuntary. The 43 percent figure is 2023 data, reported by CCAP in May 2025. Source: CCAP Attrition and Retention Survey, conducted by Willis Towers Watson across 145 member organizations. This is the one far-offshore market with a real series behind it.
• Caribbean and Latin America: no published series. There is no equivalent published attrition series for either region. Any band you see quoted for those markets, ours included, is a vendor estimate rather than a published measurement.

So the honest shape of the delta is this. The far-offshore side is measured. The Caribbean side is our estimate. Two things follow. First, treat any Caribbean attrition number, from us or from anyone else, as a claim to be tested during a pilot rather than a benchmark to plan against. Second, the part of this argument that survives scrutiny is not the number, it is the mechanism, and the mechanism is what the rest of this piece is about.

One more figure gets misused in this conversation, so it is worth pinning down. IBPAP president and CEO Jack Madrid put attrition in the wider Philippine IT-BPM industry at 30 to 40 percent in an on-the-record remark in June 2023. That is a press comment rather than a published study, and IT-BPM is broader than contact centers, so it is not a substitute for the CCAP figure above.

The mechanism: why Caribbean voice rooms hold tenure

The Caribbean attrition delta is not a marketing claim. It is the predictable output of four structural inputs that compound over a 12-month program. Any honest operator running Caribbean nearshore voice will describe the same four drivers.

1. Native-English wage parity within the local market

Jamaica, Trinidad, Belize, and the English-speaking Caribbean run native-English labor markets with voice BPO wage floors that are competitive with local market rates. Fronters are not paid at a labor-arbitrage discount relative to other local jobs they could plausibly take. When the next adjacent job pays similarly, the floor holds. When the wage floor is propped up by labor-arbitrage discount, the next adjacent job in the local economy is always poaching, and month-four attrition is structural rather than incidental.

2. US client time zone overlap without graveyard shifts

A fronter on a US Eastern shift in Kingston, Port of Spain, Belize City, or Bogota is working the same hours as US business hours, on their local UTC-4 to UTC-5 clock. Sleep is normal. Family time is normal. Weekend integrity is intact. Compare this with a far-offshore voice room in a 12-hour-offset market, where the fronter is on an inverted-circadian shift to cover US daytime. The body absorbs that cost for a quarter, sometimes two. By month nine, the floor is cycling. Time zone is not a feature, it is an attrition input.

3. Longer voice career arcs in the Caribbean

Jamaica and Trinidad have 20-plus years of voice BPO history. Kingston in particular has a multi-generational labor pool where voice work is a recognized career, not a stopgap. Fronters move from one program to the next within the same career. Supervisors and QA leads are promoted from the floor on a known ladder. The result is a tenure-friendly career architecture that does not exist in markets where voice BPO is a recent labor-arbitrage experiment.

4. Caribbean cultural norms around relationship work

Outbound voice in regulated verticals is relationship-heavy work. Debt collection under Regulation F, Medicare AEP under the CMS Medicare Communications and Marketing Guidelines, and insurance fronting under NAIC and state-licensing context all reward agents who can hold a conversation, read a consumer, and adjust tone. Caribbean cultural norms around relationship-building and conversational warmth align with what these programs reward. That alignment shows up as longer agent satisfaction curves, which shows up as lower attrition.

The compounding cost of high attrition

Attrition is a compounding cost, not a flat one. A floor that turns over half its seats in a year does not absorb that penalty once; it absorbs it every year, and each replacement cycle drags the floor's average tenure down, which drags conversion down, which drags the program's margin down.

Work an illustration, not a benchmark. Take a 20-seat outbound voice program. At the published Philippine contact center rate of 43 percent on 2023 data (CCAP Attrition and Retention Survey, conducted by Willis Towers Watson), the floor loses roughly 9 trained agents per year. Suppose a Caribbean room holds 10 points better. The same floor loses roughly 7. That 10-point assumption is ours, not a measurement, so substitute your own separations the moment you have them. The shape of the arithmetic is the point, not our number.

Then price each separation with a real cost-per-hire figure rather than a round number. Cost per hire alone averages 4,683 US dollars with a median of 1,244 US dollars (SHRM, 2022 Talent Access Report, n=472), and 5,475 US dollars for nonexecutive roles (SHRM, October 2025 Recruiting report, average). The roughly fourfold gap between mean and median is right skew, which is exactly why a single "cost to replace an agent" number hides more than it shows. Pick the one that matches your hiring model. Training, equipment and the productivity lost during ramp sit on top of cost per hire, and the size of that add-on is a Call Force Global estimate rather than a published figure. None of it shows up on the per-hour invoice. All of it shows up on the program's twelve-month P&L.

And that is just the recruit-screen-train-ramp cost. The downstream consequences are larger:

• Ramp time on conversion. We see fresh agents in regulated outbound work run at roughly 50 to 70 percent of tenured-agent conversion for the first 60 to 90 days. That band is a Call Force Global estimate from our own floors, not a published benchmark. Every separation drops conversion across the cohort while ramp completes.
• Calibration tax. QA and supervisor time spent re-calibrating fresh agents is time not spent improving tenured-agent performance. The calibration ceiling stays artificially low.
• Lost institutional context. A tenured fronter knows the buyer's product, the rebuttals that work, the regulated language that gets through compliance, and the cohort-specific patterns. Each separation is a small institutional context loss.
• Supervisor attrition cascade. Supervisors and QA leads are usually promoted from the floor. A floor that cannot hold tenure cannot grow its own supervisor pipeline.

The Caribbean attrition delta is the real Caribbean savings

Here is the insight the per-hour rate hides. The apparent Caribbean premium over far-offshore voice is something in the order of 6 US dollars per hour on the unloaded rate, which is a Call Force Global estimate from the rate cards we have seen rather than a published benchmark. For a 20-seat program at 40 hours per week and 50 weeks per year, that is 240,000 US dollars in nominal premium.

Now layer in the attrition delta. Every separation you avoid is a cost-per-hire you do not pay, plus the ramp-time conversion drag on the replacement, plus the calibration tax, plus the lost institutional context. Two or three avoided separations a year on a 20-seat floor already moves five figures once you use your own cost per hire. For verticals where every agent-week of tenure is worth measurable conversion dollars (Medicare AEP, debt collection, insurance fronting), the operational gap can offset a meaningful share of the nominal premium. We are deliberately not giving you a single headline savings figure, because the honest version of it depends on numbers only you have.

This is why senior procurement teams who have run both programs tend to reach the same conclusion. On a TCO basis the headline 6 dollar Caribbean premium is not really a premium. It is the price of buying out of a churn problem that the per-hour quote refuses to name.

The Caribbean attrition delta is the gap between published far-offshore contact center attrition, the Philippines at 43 percent on 2023 data (CCAP with Willis Towers Watson, reported May 2025), and Caribbean nearshore attrition, for which no published series exists. Same-time-zone scheduling, native-English career perception, and smaller competing labor pools are the structural reasons we expect the Caribbean side to run lower. The Caribbean figure itself is a Call Force Global estimate, not a measurement.

Verticals that live or die on agent tenure

The Caribbean attrition delta matters more in some verticals than others. For unregulated outbound (cold lead-gen, low-stakes appointment setting), a high-churn floor will still produce volume even if conversion suffers. For regulated outbound where the same agent talks to the same buyer multiple times, tenure is the program. Three verticals where the delta is most load-bearing in 2026:

Medicare AEP front-end work

The CMS Medicare Communications and Marketing Guidelines require disclosure language that a tenured fronter internalizes over weeks of reps. Fresh agents in an AEP rotation are a compliance liability and a conversion liability simultaneously. A program that cycles agents through AEP is a program that re-trains compliance language in October every year, which is exactly when it cannot afford to.

Debt collection under Regulation F

Reg F nuance is built up over months of warm-transfer reps. The agent's judgment on disclosure timing, the agent's read on consumer hardship, the agent's ability to maintain relationship integrity across multiple touchpoints, all depend on tenure. On a debt floor, every point of annualized attrition is a matching point of regulated-context expertise walking out the door.

Insurance fronting

Insurance fronters do not need to be licensed (the licensed work sits with the client's US agents on the receiving end of the warm transfer), but they do need NAIC and state-licensing context to pre-qualify cleanly. That context is built up over months. A high-attrition insurance fronter floor is a floor where the pre-qualification quality drops every quarter as the cohort cycles.

For these verticals, the Caribbean attrition delta is not a nice-to-have. It is the program.

CFG's posture on attrition

Call Force Global does not maintain an audited attrition dataset and does not publish an attrition percentage. We used to point at an industry benchmark to imply one, and that was not a defensible thing to do, so we stopped. What we believe, based on running native-English fronter rooms in Jamaica, Trinidad, Belize, and Colombia, is that these floors hold tenure better than a graveyard-shift far-offshore floor. That is a Call Force Global estimate, not a measurement. We share program-specific separations under NDA during pilot scoping, which is where you should be checking it anyway.

The structural reason behind our delta is the same as the structural reason behind any honest Caribbean operator's delta: native-English wage parity, US Eastern shift overlap, longer voice career arcs in our locations, and cultural fit for relationship work. We do not import a far-offshore labor-arbitrage wage model into the Caribbean. We pay competitive wages relative to local market rates, and the floor holds.

One additional structural input on the CFG side: our named-bench replacement model (see the homepage explanation of how CFG covers seat continuity) means that when a fronter does leave, the seat is held by a named bench agent who has already trained on the client's program. The client does not experience the calibration tax even on the separations that do occur.

How to measure the attrition delta in your own contract

Procurement teams can run this in five steps without any proprietary data:

1. Calculate your program's annualized voluntary attrition. Separations per 12 months divided by average headcount. If your current vendor will not share this number, that is itself a finding.
2. Compare against the published Philippine figure if your current floor is far-offshore. Contact center attrition there was 45 percent in total in 2022 and 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). If you are well above it, your program is underperforming its own market.
3. Compare against the published US figure if you are weighing onshore. A 27 percent mean and a 21 percent median, with a third of respondents above 30 percent (ContactBabel, US Contact Center Decision-Makers' Guide, 2024 edition, 189 US contact center managers). There is no published Caribbean or Latin American series to compare against, so for the nearshore option you are testing a vendor's estimate rather than checking a benchmark.
4. Multiply each separation by your own loaded cost per hire. If you do not have one, SHRM puts cost per hire at a 4,683 US dollar average and a 1,244 US dollar median (2022 Talent Access Report, n=472) and 5,475 US dollars for nonexecutive roles (October 2025 Recruiting report, average). Add your own training, equipment and ramp cost on top. Express the result in dollars per year, not percentage points.
5. Add ramp-time conversion drag. For each separation, assume the replacement runs at a fraction of tenured-agent conversion for 60 to 90 days. Our own floors suggest 50 to 70 percent, which is an estimate rather than a published figure, so use your own ramp curve if you have one. Multiply by your program's daily revenue per seat. This is the conversion side of the delta.

Run with your own inputs, this calculation usually produces a five-figure annual number for programs in the 10 to 50 seat range, before counting the supervisor pipeline effect or the institutional context loss. We are not going to quote you a range, because the range depends entirely on your separations and your cost per hire, and a vendor-supplied range is exactly the kind of number this piece is arguing against.

Conclusion

The Caribbean attrition delta is a real argument, but it is half measured and half estimated, and pretending otherwise is how this topic got polluted in the first place. The far-offshore side has a published series behind it. The Caribbean side does not, and no vendor has one. What does hold up is the mechanism: wage parity, shift overlap, career arc, and cultural fit. For Medicare AEP, debt collection, and insurance fronting, tenure is the program. For procurement teams modeling 2026 voice programs, the delta is the answer to why the per-hour quote is the wrong number to optimize, as long as you run it on your own numbers.

CFG runs fronter-only rooms in Jamaica, Trinidad, Belize, and Colombia. We pre-qualify, we do not close, and we warm-transfer regulated work to the client's licensed US agents. The CFG outsourcing calculator models the delta against your current program's numbers. Or get my 24-hour quote and we will run the five-step calculation against your actual attrition data under NDA.

For related context, see the Caribbean Fronter Cost Curve methodology for the five-component cost composition framework, and the 2026 nearshore call center vendor list for a side-by-side of operators in the same wedge.

Sources

• ContactBabel. US Contact Center Decision-Makers' Guide, 2024 edition, 189 US contact center managers. US contact center attrition at a 27 percent mean and a 21 percent median, with a third of respondents above 30 percent. US-only, segmented by vertical and company size, not by geography.
• ContactBabel. UK Contact Center Decision-Makers' Guide. UK contact center attrition of 15 to 32 percent across 2003 to 2023, recently 23 to 24 percent.
• Contact Center Association of the Philippines (CCAP). Attrition and Retention Survey, conducted by Willis Towers Watson across 145 member organizations. Philippine contact center attrition of 45 percent in total in 2022 (31 percent voluntary, 14.1 percent involuntary) and 43 percent on 2023 data, reported May 2025.
• SHRM. 2022 Talent Access Report (n=472): cost per hire of 4,683 US dollars average, 1,244 US dollars median. October 2025 Recruiting report: 5,475 US dollars average for nonexecutive roles.
• US Bureau of Labor Statistics. Occupational Employment and Wage Statistics, SOC 43-4051 Customer Service Representatives, May 2025: 21.53 US dollars per hour, 44,770 US dollars per year. Wages only. BLS publishes no separations rate for this occupation.
• Correction, August 2026. Earlier versions of this page built the comparison on two attrition bands, one presented as a far-offshore figure and one as a global industry average, each attributed to a named industry source. Neither attribution held up when we went back and checked the underlying documents. Both bands have been retracted and replaced with the published figures listed above.
• Centers for Medicare and Medicaid Services. Medicare Communications and Marketing Guidelines (MCMG). Annual editions.
• Consumer Financial Protection Bureau. Regulation F, debt collection practices final rule.

Frequently Asked Questions