The Talent You Lose at Month 18 — and Why It's Rarely About Salary

Most companies believe the battle for talent is won or lost during hiring. In reality, it’s often lost midway — around the 18-month mark, when an employee reaches peak productivity. And surprisingly, it’s rarely about salary.

When does hard-won talent actually leave?

There’s a predictable window: between 18 and 24 months of tenure. By then, the employee understands the code, knows the system, has delivered significant projects, and the “honeymoon” phase of the new job is over. This is precisely when they are most valuable — and most likely to leave.

The underlying data backs it up. According to the Stack Overflow Developer Survey 2025, only 24% of developers report being happy at work (up from 20% the previous year), and just 46% say they are not looking for a job. In other words, more than half of your engineers are, to some degree, considering leaving. The average tenure for a developer is around two years, with roughly 69% having been at their current company for less than two years.

When we talk about “churn,” we mean turnover — the percentage of people leaving who need to be replaced. The problem isn’t that churn exists (it always will). The problem is that most companies focus on it in the wrong place.

Why do they leave if they’re well-paid?

Here’s the uncomfortable part. The same Stack Overflow 2025 survey shows the top three factors influencing a developer’s satisfaction are, in order: autonomy and trust, competitive compensation, and solving real problems. “Getting along with the manager” ranked ninth.

That reorders priorities differently than most boards assume. Salary matters — it’s second — but it’s not the trigger. What drives someone to leave around month 20 is rarely low pay; it’s the lack of growth. Retention analyses from 2026 and Gartner data agree that the absence of future development is the number one reason for leaving, cited by roughly 40% of those who resign.

There’s a second, quieter factor: salary compression. Someone hired at market rate 18 months ago, if never adjusted, is likely 10 to 20% below what they could earn elsewhere today. They don’t leave out of greed. They leave because the market moved and you didn’t. When an external offer arrives, the number merely confirms a decision they were already leaning toward.

How much does it cost to lose someone at month 20?

Far more than any spreadsheet shows. SHRM estimates replacing an employee costs between 6 and 9 months of their salary. Gallup places it between 50% and 200% of annual salary depending on the level — and for technical roles, around 80% of salary. Add the 8 to 12 weeks it takes to hire a senior, plus the ramp-up months until the replacement performs at the same level.

But the most expensive cost isn’t visible. It’s the knowledge that walks out with the person: how that undocumented service works, why a particular architectural decision was made, who’s who on the team. You can’t recover that by hiring quickly.

Stage	Typical cost	What is lost
Month 0–3 (onboarding)	Low	Cultural friction, misaligned expectations
Month 18–24 (mid-tenure)	6–9 months of salary	Full productivity + undocumented knowledge
Replacement	8–12 weeks search + ramp-up	Business context, team relationships

The irony is that month 20 is the most expensive to lose and the least managed. Everyone focuses on onboarding; almost no one manages the point where the investment finally starts to pay off.

Why does no one see the departure coming?

Because most companies measure performance too late and satisfaction never. The resignation at month 20 isn’t a bolt from the blue — it’s the end of a story that started months earlier, with small signals no one was watching. Less enthusiasm in meetings, less initiative, a personal project that quietly fizzles out.

What we observe in practice — after more than 500 hiring processes across LATAM and monitoring how people evolve once inside — is that the real risk isn’t at the beginning. It’s around month 18, and almost no one is measuring it. By the time a manager notices someone “seems off,” it’s usually too late: the decision is made.

You don’t need a complex system to catch it. A light, constant pulse is enough: a brief but honest conversation each month, a check on whether the person is still growing or on autopilot, and a compensation review that doesn’t wait for a counteroffer. Retention isn’t fixed with an emergency raise the day someone announces they’re leaving. It’s fixed by looking ahead.

What this means for your company

If your talent strategy ends the day the person signs, you’re focusing on the cheaper part of the problem. Hiring is the entry. Retaining someone productive at month 20 is where the real return on your investment lives.

The important question isn’t “how do I hire faster.” It’s “who on my team is approaching the 18-month window, and what do I actually know about how they’re doing?” If the answer is “I don’t know,” that’s the work. Talent isn’t lost in the search. It’s lost when you stop paying attention right after it starts to deliver.