First-Year Attrition

Because it is the earliest point where a hiring decision can be judged by an outcome rather than an impression. Everything recruiting measures before day one is a proxy: interview scores, references, assessment results. The first year is where those proxies get tested against the actual job. A cohort that leaves early tells you the process selected people who could not or did not want to do the work as it really is. That said, recruiting does not own the whole number. Onboarding, the manager, team stability and how honestly the role was described all contribute. Read it as a shared signal, and split it by hiring manager and by source before assigning any cause.

Overall attrition covers everyone in the company, so it is dominated by long-tenured staff and moves slowly. A stable overall rate can sit on top of a serious new-hire problem and show nothing. Probation failure is narrower than first-year attrition: it captures only formal terminations inside a defined probation window, missing the person who passed probation and resigned in month eight. First-year attrition sits between the two and is the one that maps to hiring decisions. Track all three if you can, but do not substitute one for another in a report. Also decide whether internal transfers count as leaving. Most teams exclude them, since an internal move usually means the hire worked, just not in that seat.

Cohort size, first. A team that hired six people last year has a rate that jumps enormously when one leaves, and reading that as a trend wastes everyone's time. Seasonal bulges create a second problem: a large intake in one month produces a wave of departures roughly a year later that looks like a sudden deterioration and is really the same cohort maturing. Mixing voluntary and involuntary exits is the third distortion, and it matters because they mean opposite things. Someone let go for performance points at selection; someone who resigned may point at the manager, the commute, or a better offer. Split the number by exit type, by cohort month, and by manager before drawing a conclusion, and quote raw counts beside percentages.

Go to the exit reasons before touching the hiring process. If leavers describe a job that differed from what they were told, the fix is honesty in the posting and the interview, including the parts candidates find unattractive. If they describe a manager problem, that belongs to the manager's own development rather than to sourcing. If they describe being unable to do the work, the assessment stage was not testing the real job, and a work sample usually fixes more than another conversation round. Structured interview question sets and written scorecards make the diagnosis possible at all, because unstructured interviews leave no record to review. A hiring platform that keeps assessment evidence with the hire record turns this into an analysis rather than a memory exercise.