By the time the resignation letter lands, the data has flagged the risk for months. Rising unplanned absence is the most reliable leading indicator of voluntary attrition, and the most ignored. This piece shows you why the signal works, and why the signal sits unread in a spreadsheet at most Australian SMBs.
Engagement does not collapse on the day someone resigns. Engagement fades quietly. Attendance is the first place the fade turns visible and measurable. People stop showing up in spirit before they stop showing up in person. The days off are the early record.
Most retention tools lag. An exit interview tells you why someone left after they have gone. An annual engagement survey gives you a snapshot, stale within weeks and easy to game. Performance dips show up late, once disengagement has already cut output.
Unplanned absence works differently. Absence leads because absence is behavioural, not self-reported, and absence changes early. The mechanism is plain. As someone disengages through burnout, a poor manager relationship, or feeling overlooked, the bar for taking a day off drops. A scratchy throat once pushed aside becomes a sick day. The pattern shifts before the person has decided to leave.
Tells you why, after the chance to act has gone.
A periodic snapshot, easy to game, stale within weeks.
Behavioural, continuous, and shifts months ahead. Already in your data.
The total number of days rarely matters. The change in shape matters. Shape shows up only when you plot the data instead of listing it.
Single days, more often, with vaguer reasons. Frequency beats duration as a signal. One long, certified illness is usually an illness.
A drift toward long weekends signals a person checking out, or a life issue a conversation resolves. You want to see either one.
The reliable person who suddenly drops off tells you more than someone who has always taken their full allowance. Context beats raw counts.
The insight is hard money, not a soft idea. Replacing a mid-level employee costs 50 to 200 percent of annual salary once you count recruitment, lost productivity, and ramp-up.¹ For an Australian SMB, one avoidable resignation erases a meaningful share of the year's margin. The absence pattern gave you the cheap, early warning, the version of the problem you still had time to fix.
A second cost follows. Every unmanaged absence lands on the people who cover. The colleagues who pick up the slack week after week grow more likely to disengage. One departure sets up the next.
The goal is not surveillance. The goal is noticing in time to care. Three principles keep you on the right side:
Absence turns scattered sick days into day-of-week patterns and a return-to-work queue. The signal reaches you months before the exit interview does. Free for teams of 5.
Workforce research supports the link between absence patterns and attrition, but the link is probabilistic, not deterministic. A pattern prompts you to understand an individual's situation. A pattern does not predict any one person. Figures are point estimates and vary by sector and method.
