A firm of eleven people loses the same position three times in two years. Each departure has its own explanation: the first person moved out of state, the second was not quite up to the pace, the third left for a competitor offering more money. Every explanation is true and none of them is the reason, because a seat that empties three times while every other seat holds is telling you something about the seat. This composite, assembled from a pattern common in small businesses, is what that usually turns out to be.
The Role as It Was Written and as It Was Lived
The position was created when the owner could no longer do everything, and it was defined by subtraction: the tasks the owner least wanted, gathered into one job. Scheduling and dispatch, customer calls, ordering, invoicing, chasing payment, and the paperwork for two regulatory filings nobody understood. On paper it read as an office coordinator role. In practice it was three jobs held together by whoever occupied the chair, and its boundaries changed whenever the owner had a busy week.
It also had a structural feature nobody named: the person in it was the only one who did any of it. There was no second person to hand anything to, no coverage during vacation, and no one who could answer a question about how something worked. The role was simultaneously the busiest in the building and the least supported, which is a combination that produces exactly the outcome the firm kept experiencing.
What the Exit Conversations Actually Said
The firm did hold exit conversations, which puts it ahead of most, and each one produced a comment the owner filed as personal rather than structural. The first person said she never felt like she had finished anything. The second said he was interrupted constantly and could not tell which of the interruptions were more important than what he was doing. The third said, when pressed, that the pay was a factor but the real issue was that nobody ever knew whether the job was being done well.
Read individually those are three different complaints from three different people. Read together they are one complaint stated three ways, and it is not about pay or pace at all. It is that the role had no priority order, no boundaries, and no feedback, so the person occupying it could work hard all day without ever being able to say whether the day had gone well. That is exhausting in a way that is difficult to articulate at an exit interview, which is why it comes out as something else.
Why It Kept Happening
Each time the seat emptied, the firm ran the same process again and concluded it had hired the wrong person. The advertisement was reposted, references were checked more carefully, and the interview added a question about handling pressure. Every one of those responses treats the problem as selection, which is the natural conclusion when three different people fail at the same thing and the only variable anybody can see is the people.
The cost of that misreading is larger than it looks. Three recruitment cycles, three onboarding periods during which the work is done badly or not at all, three windows where the owner absorbs the tasks personally, and a slow reputational effect in a small town where people talk about which employers churn. Add the accumulated knowledge that walked out three times, and the cost of not examining the seat is a multiple of what fixing it would have taken.
What the Fourth Attempt Did Differently
Before advertising again, the owner spent a day writing down what the role actually contained, in hours, and the exercise answered the question immediately: the list ran to about sixty hours of work a week. That is not a job anybody can hold, and it explained every departure without reference to any individual. Two of the regulatory filings moved to an external bookkeeper, the ordering moved to a technician who was already handling half of it informally, and what remained was a genuine forty hour role.
Three other changes came out of the same day. A written priority order stating what gets dropped when everything cannot be done, which is the single thing the previous three occupants most needed and never had. A rule that interruptions during the invoicing block go to the owner instead. And a standing fifteen minutes on a Friday to say what had gone well and what had not, which is the cheapest form of feedback available and the one small employers most consistently skip.
How to Tell Whether It Is the Seat or the People
The test is arithmetic rather than intuition. Write down every task in the role and estimate the hours honestly, including the ones nobody counts, then compare the total to the hours the position is paid for. If the answer exceeds the paid hours by any meaningful margin, the role is the problem and no hire will fix it. If the hours fit, look next at whether the person can state their own priority order, and whether anybody has told them in the last month how they are doing.
Job openings and separations are counted separately in the national figures the Bureau of Labor Statistics compiles, and one useful habit borrowed from that framing is to track your own turnover by seat rather than by headcount. A business with steady overall numbers can be losing one position repeatedly while everything else holds, and that pattern is invisible in an annual total. The fourth person is still in the job, which is the outcome, and the thing that changed was never a hiring decision.