A stalling company does not lose talent randomly. The pattern of departures is non-uniform, it is predictable, and it precedes the deterioration in financial performance by something in the range of twelve to eighteen months. The first people to leave are the people with the most external options, which happen to be the same people who were doing a disproportionate share of the work. Their departure is a signal. It is also, typically, the earliest signal a company has that the business trajectory has changed.
This is not what most executive teams believe is happening. When a top performer leaves, the exit is usually categorized as idiosyncratic: better offer, family reasons, a new challenge. A second departure a month later is coincidence. A third is a wake-up call, but only if someone is counting in a specific way. The aggregate signal (that the top quartile of performers has started moving at a noticeably higher rate than the rest of the population) is rarely visible to the people who need to see it, because companies do not measure attrition by performance cohort.
When they do measure it, the pattern is striking. Top performers in a healthy company leave at roughly half the rate of the broader population. In a deteriorating company, that ratio inverts before the deterioration is visible in any other metric.
- Top performers in high-growth companies have voluntary attrition rates of 6 to 8 percent, versus 12 to 15 percent for the broader employee base (Gartner, 2023).
- In the 18 months preceding a significant performance deterioration, top performer attrition has historically risen 1.6x to 2.2x, while broader attrition remains stable (LinkedIn Workforce Insights, cross-industry sample).
- The cost of replacing a top performer is estimated at 2.0x to 2.5x annual compensation, once onboarding, productivity ramp, and institutional knowledge loss are factored in (SHRM, 2024).
The 18-month lead time shown in Figure 1 is not a coincidence. It is a function of how A-players respond to signals about the future of a business. The high performers in any organization are also, almost by definition, the best-connected. They see trends earlier, they have more external options, and they make career decisions on longer time horizons than the median employee. When they begin to sense that the trajectory has changed, they do not wait for the financial proof.
The selection pressure inside a slowing company
The mechanism has less to do with dissatisfaction than with option value. Every employee of every company is making a continuous, largely unconscious bet on the equity of their time. The question is: what is the expected value of another year spent here, compared to the expected value of another year spent somewhere else. The calculation incorporates compensation, career trajectory, the quality of the work, the quality of the team, and the option value of association with the company's future.
For a median performer, the calculation is fairly stable. They have moderate options, they face moderate friction to move, and the marginal company trajectory has moderate effect on their equation. They stay through a lot of volatility. For a top performer, the calculation is much more sensitive. They have many options. The friction to move is low (recruiters call them). The trajectory of their current employer has a large weight in their equation, because association with a rising company is worth a great deal and association with a declining one is worth very little.
This is why A-player attrition is a leading indicator and broad attrition is a lagging one. The A-player is running a more precise version of the same calculation everyone else is running, and they are running it against better information. When they start leaving, they are not reacting to the miss that has already happened. They are reacting to the miss that has not happened yet.
Inside a deteriorating company, the selection pressure compounds. Each A-player departure makes the remaining A-players slightly more likely to leave, because the quality of their environment (their peers, the work they get access to, the information flow) has just dropped. The remaining top talent is now carrying a higher proportion of the critical work, which accelerates burnout, which accelerates their own calculation. The company that loses two A-players in a quarter is substantially more likely to lose two more in the next quarter than a company that lost none.
The cost in Figure 2 is conservative. It excludes the second-order effects that are harder to quantify but usually larger: the signal the departure sends to remaining top performers, the capacity gap while the role is open, the degradation of work quality during a six-month productivity ramp, and the compounding cost when one A-player exit triggers two more.
By the time the exit is logged as a resignation, the cost has already been paid. The question is whether anyone in the company is counting it.
Why companies miss the signal
Three structural reasons explain why this pattern, despite being consistent and well-documented, surprises executive teams each time it unfolds.
The first is that most companies do not segment attrition by performance tier. Aggregate voluntary attrition is reported, sometimes broken out by function or level. Performance-weighted attrition almost never is. If your 12 percent attrition is concentrated in the bottom quartile, that is a different company from one where 12 percent is concentrated in the top quartile. The HR dashboard treats them identically.
The second is that individual A-player departures are absorbed rather than investigated. When a top performer leaves, the HRBP conducts an exit interview. The exit interview captures the employee's stated reason. Stated reasons are usually sanitized versions of the real ones. The honest answer ("I don't think this company is going anywhere") is almost never given, because the departing employee has nothing to gain by giving it and something to lose. The recorded data is therefore systematically biased toward benign explanations.
The third is that the CEO, by the time they hear about a specific A-player exit, hears about it framed as a win for the departing person rather than a loss for the company. "She got an amazing opportunity at X." The narrative is inoffensive and it prevents the aggregation that would reveal the pattern. The CEO sees eight good-news departures over twelve months and does not connect them.
What to do with the signal
The diagnostic value of A-player attrition lies in what happens when you start measuring it correctly.
Three practices, in sequence, change what a company can see. First, segment attrition by performance quartile and review it monthly, not annually. Second, calibrate the ratio: if your top quartile attrition has risen above your bottom quartile attrition, you are in a selection event, regardless of what the headline number says. Third, when the ratio inverts, treat it as a business trajectory question, not an HR question. Ask what the A-players are seeing that the leadership team is not.
The answer is almost always instructive. Sometimes it is external (the market is shifting, a competitor is pulling talent). More often it is internal: decision speed has slowed, the quality of work has declined, senior leadership has stopped developing people. These are the structural issues this publication exists to surface, and A-player attrition is one of the earliest and most honest signals a company has about whether they are taking hold.
The best talent is paid to read the future accurately. When they leave, they are not quitting the company. They are pricing it.