The number nobody wants to publish: across our 500 placements between 2023 and 2025, year-one churn for LATAM hires landed at 32% blended. Mexico ran 24%, Colombia 28%, Argentina 38%, Chile 26%, Brazil 30%. These figures are from Selection Book's internal placement panel and may not match industry-wide LATAM churn (no public dataset isolates year-one churn by country at this granularity). Public benchmarks for total annual tech attrition run 13-21% — our LATAM-specific year-one number sits well above that, consistent with the cross-border friction this article unpacks. Every percentage point above 20% is preventable. Most US scale-ups don't try.
The cost math: replacing a $75K mid-level engineer in Mexico costs $60,000-$90,000 on the full-loaded view (1-1.5x annual base) — recruiter fee (10-15% annual comp), 3-6 months ramp at 50% productivity, manager hours spent recovering, and the project-level cost of context loss. Direct hard costs alone (recruiter + onboarding) run closer to $40,000-$45,000 — the rest is opportunity cost most CFOs don't book. Most companies could spend half the full-loaded number on retention and save the other half.
The five drivers of year-one churn
| Driver | When it hits | % of departures it explains |
|---|---|---|
| 1. Under-market comp adjustment | Month 9-14 | ~35% |
| 2. No visible career path | Month 5-8 | ~22% |
| 3. Decision-isolation from US team | Month 3-7 | ~14% |
| 4. Missed promotion velocity | Month 10-14 | ~13% |
| 5. Tax-side surprise (first March filing) | Month 4-6 | ~7% |
| Other / personal | Variable | ~9% |
Driver 1 — Under-market comp adjustment (month 9-14)
The pattern: you hire at $68K. Six months in, the LATAM senior backend market moves to $74-78K (it did in 2024 and 2025). Your employee gets approached by a competing US scale-up at $76K. You counter at $72K because that's your band. They leave.
The math is brutal: bumping someone from $68K to $76K costs $8K/year. Replacing them costs $60-90K fully loaded ($40-45K direct hard costs alone). The ratio is 5-11x.
The fix: proactive comp reviews at month 9 (not month 12) using fresh market data. Bring the bump before they have the competing offer. Once they have the offer, you're negotiating from weakness and accept rate drops to 35-40% even at-market.
Driver 2 — No visible career path (month 5-8)
The pattern: a senior engineer in Mexico City has watched two US peers get promoted in the time she's been at your company. She hasn't had a single conversation about her own trajectory. By month 6 she's quietly interviewing.
This is a US-management failure imported to LATAM. US engineers tolerate ambiguous career paths because they have informal signals (cross-functional projects, mentor relationships, lunchroom conversations). Remote LATAM engineers have none of those signals. They need explicit ones.
The fix: career-ladder document published in month 2. Levels, expectations per level, comp band per level. Quarterly career conversations on the calendar (not "as needed"). Promotion velocity benchmark stated explicitly: "typical IC3 → IC4 is 18-24 months at this company."
Driver 3 — Decision-isolation from US team (month 3-7)
The pattern: architectural decision gets made in a hallway conversation between two US engineers. Mexico-based engineer reads about it in a Slack channel afterward and realizes she should have been in that conversation. Six months of this and she concludes the company doesn't actually value her input. She leaves.
This is the hardest driver to fix because it's structural, not procedural. Remote isolation is the default state — you fight against it actively, or it wins.
The fix: all architectural decisions documented as written RFCs in a shared doc, with explicit comment period (3-5 business days). Hallway decisions are explicitly disallowed. Project kickoffs happen with all stakeholders in the same Zoom, not a US-only meeting followed by an async update.
Driver 4 — Missed promotion velocity (month 10-14)
Closely related to Driver 2, but different. Driver 2 is about whether there's a path. Driver 4 is about whether the path moves as fast as US peers.
The data: US engineers at scale-ups average 20-26 months for IC3 → IC4. LATAM engineers at the same companies average 28-36 months for the same transition. That gap of 6-12 months is silent and unjustified. It's almost always just rater bias — US managers see US engineers more often and credit them more readily.
The fix: calibration sessions for promotions that explicitly compare LATAM and US engineers head-to-head. Track promotion-velocity ratio (LATAM promotion rate / US promotion rate) as a KPI. Anything below 0.85 means you have a systematic bias.
Driver 5 — Tax surprise in first March (month 4-6 if hired in fall)
The pattern: engineer was hired in November 2024 as a Mexican contractor receiving USD via Wise. In March 2025 they file their first Mexican tax return and discover they owe 30-35% of receipts they didn't withhold. Cash crunch, anxiety, looking for a "safer" employer.
The fix is small but high-leverage: if you're paying as contractor, hand them a Mexican accountant for the first year ($150-$300/month). If you're paying via EOR, this driver doesn't exist (taxes are withheld). For most US scale-ups, the lesson is: don't use contractor structure for someone who will be exposed to tax surprise in March.
What works — by phase
Career-ladder published, comp band visible, 1:1 cadence set
Hand them three documents in week 1: career ladder with comp bands, the company's promotion rubric, and a 6-month onboarding plan with check-in dates. This sounds basic. Almost nobody does it. The ones that do see 18-24 month churn drop to 12-14%.
Fresh benchmark + comp conversation before the market does it for you
At month 8, pull fresh LATAM salary data. If the market has moved 5% or more, adjust proactively. Tell the employee what you're doing and why. The pre-emptive bump buys 2-3x the loyalty of the same bump after they've gotten a competing offer.
Explicit promotion path, with calibration against US peers
At month 9-12, hold the explicit "where are you on the ladder" conversation. Use specific behavioral examples. Calibrate against a US peer at the same level — same expectations, same comp band, same velocity.
What the data says about who stays
| Retention factor | Effect on 18-mo retention |
|---|---|
| Pre-emptive comp adjustment at month 9 | +17 pts |
| Career-ladder doc published in month 2 | +12 pts |
| Quarterly in-person team gathering | +10 pts |
| Direct manager based in similar time zone | +9 pts |
| EOR (vs contractor) with statutory benefits | +8 pts |
| Equity grant (any form, even phantom) | +6 pts |
| In-person kickoff visit in first 90 days | +5 pts |
Each driver is small. The compound is large: a company that does the top 4 plays (proactive comp, career ladder, in-person quarterly, EOR not contractor) sees 18-month retention of 78-84% vs market-average 62-68%. That delta is roughly $32,000 per hire per year in avoided replacement cost.
Key takeaways
- LATAM remote churn averages 24-38% year one. Half of departures are preventable.
- Comp adjustment at month 9 is the highest-leverage retention play. Pre-empt the market, don't wait for the competing offer.
- Publish the career ladder in month 2. Most US companies don't, and it's the second-largest driver of departures.
- Track promotion-velocity ratio (LATAM / US). Anything below 0.85 means rater bias and slow-burn churn.
- Use EOR not contractor for long-term hires. The +8 pts retention is worth more than the EOR fee.