The workforce analytics metrics that matter most to operations teams are absence rate, Bradford Factor, lateness rate, overtime percentage, labour cost percentage, scheduled versus actual hours variance, shift fill rate, schedule adherence, agency share, turnover, early-leaver turnover and off-schedule access events. Together they explain cost, coverage and risk.
A workforce analytics metric is a repeatable calculation on workforce records, such as clock-ins, rosters, leave, payroll hours and access logs, that tells you whether people, time and money are being used as planned. Many published lists focus on recruitment and engagement. This one is for teams that run shifts, often across several sites, and need to know where hours and cost are going this week, not just at year end.
The 12 metrics at a glance
Use the table as a reference. Each metric has a formula you can run in a spreadsheet, a reason to care and a warning sign that should prompt a closer look.
| # | Metric | Formula | Why it matters | Warning sign |
|---|---|---|---|---|
| 1 | Absence rate | Hours lost to absence ÷ scheduled hours × 100 | Shows the share of planned time you did not get | Rising for 3 months in a row, or one team well above the rest |
| 2 | Bradford Factor | S² × D (spells squared × total days) | Weights frequent short absences more than one long one | Scores clustering in one shift or under one manager |
| 3 | Lateness rate | Late starts ÷ scheduled starts × 100 | Late starts delay handovers and first output | Lateness concentrated on one shift start time |
| 4 | Overtime percentage | Overtime hours ÷ total hours worked × 100 | Overtime is usually paid at a premium and hides gaps | The same people on overtime every week |
| 5 | Labour cost percentage | Total labour cost ÷ revenue × 100 | Connects workforce decisions to margin | Rising while revenue is flat |
| 6 | Scheduled vs actual variance | (Actual paid hours − scheduled hours) ÷ scheduled hours × 100 | Finds paid time nobody planned | Consistent positive variance from early clock-ins or late finishes |
| 7 | Shift fill rate | Shifts filled to required headcount ÷ shifts scheduled × 100 | The plainest measure of coverage | Weekend or night shifts regularly below 95% of required cover |
| 8 | Schedule adherence | Minutes worked as scheduled ÷ scheduled minutes × 100 | Shows whether rosters reflect reality | Large gaps between roster and clock data on specific days |
| 9 | Agency and temp share | Agency hours ÷ total hours worked × 100 | Agency cover is flexible but usually costs more per hour | Agency hours standing in for a vacancy that has been open for months |
| 10 | Turnover rate | Leavers in period ÷ average headcount × 100 | Every leaver means recruitment, training and cover cost | One site or role with turnover far above the average |
| 11 | Early-leaver turnover | Leavers with under 6 months' service ÷ starters in the same period × 100 | Points to onboarding, rota or role-fit problems | New starters leaving after their first rota cycle |
| 12 | Off-schedule access events | Building entries with no matching shift or approved visit ÷ total entries × 100 | Links security risk to workforce records | Entries by leavers, or regular entries hours before a shift |
How to calculate the absence and attendance metrics
Absence rate is the anchor metric. Count absence in hours, not days, if your shifts vary in length, otherwise a missed 12-hour night shift counts the same as a missed 4-hour evening. For a national reference point, the Office for National Statistics reported that the UK sickness absence rate was 2.0% of working hours in 2024. Your rate will differ by sector, shift pattern and how you record absence, so compare against your own trend first.
The Bradford Factor adds a frequency view. Eight one-day absences score 8² × 8 = 512, while one eight-day absence scores 1² × 8 = 8. Use it as a prompt for a conversation, never as an automatic trigger, and exclude disability-related and protected absences. Our absence rate calculator and Bradford Factor calculator run both formulas for you.
Lateness rate needs a clear rule: decide your grace period, apply it the same way everywhere and report trends rather than single incidents. More on reading absence alongside other data is in attendance and absence.
How to measure labour cost and overtime
Overtime percentage, labour cost percentage, scheduled versus actual variance and agency share all answer one question: are we paying for hours we did not plan? Read them together. A rise in overtime with a falling shift fill rate usually means a vacancy or absence gap. A rise in overtime with a full roster often means shift design or end-of-shift handovers are running over.
Scheduled versus actual variance is the one most teams skip. It catches early clock-ins that are paid, late finishes nobody approved and rounding rules that drift in one direction. The overtime cost calculator and labour cost percentage calculator help you put a number on each. See how these fit together in labour cost.
Worked example: one site, one month
The figures below are illustrative, not benchmarks. A warehouse runs 40 people on two shifts. In October it scheduled 6,400 hours.
- Hours lost to absence: 224. Absence rate = 224 ÷ 6,400 × 100 = 3.5%.
- Overtime hours: 480 out of 6,720 hours worked. Overtime percentage = 7.1%.
- Actual paid hours were 6,720 against 6,400 scheduled. Variance = 320 ÷ 6,400 × 100 = 5.0%.
- Night shifts filled to required headcount: 38 of 44. Night shift fill rate = 86%.
- Agency hours: 400. Agency share = 400 ÷ 6,720 × 100 = 6.0%.
Read alone, each number is a concern. Read together, they tell one story: nights are short, the gap is being filled by overtime and agency, and the 5% variance is mostly that overtime. The fix is a night shift coverage problem, not a general overtime crackdown.
How to measure coverage and turnover
Shift fill rate and schedule adherence are your coverage pair. Fill rate says whether enough people were rostered; adherence says whether they were there when the roster said. Track both by shift, not just by week, because a weekly average can hide a night shift that is short every Sunday. Our guide to scheduling and coverage covers this in more depth.
Turnover and early-leaver turnover connect coverage to cost. When early-leaver turnover is high, look at the first rota new starters receive: frequent nights or split shifts in the first month are a common reason people do not stay.
Why access events belong in workforce analytics
Off-schedule access events sit outside most HR reporting, but they belong here. An entry badge used by someone who left last month, or entries that never match a shift, are workforce data problems as much as security ones. Comparing access logs with rosters and leaver lists turns a raw log into a short exception list. If your doors run on a cloud access control system, entry events are already timestamped and tied to a person, which makes this comparison straightforward. The first article in this series, What is workforce intelligence?, explains why these sources are worth joining up.
How to start tracking without a big project
- Pick four metrics: absence rate, overtime percentage, shift fill rate and scheduled versus actual variance.
- Agree one definition for each, in writing, so every site counts the same way.
- Report them weekly, by site and shift, with a 13-week trend line.
- Set one warning threshold per metric and review only the exceptions.
- Add the remaining metrics once the first four are trusted.
Next step
If your figures live in separate attendance, rota, HR and payroll systems, the slow part is joining them every week. TeamzMate reads those records continuously, flags the exceptions that need a person and explains what changed in plain language, while decisions stay with your managers. Start with the free workforce calculators, or book a demo to see the 12 metrics on your own data.