The Bradford Factor is a formula that scores employee absence by weighting frequency far more heavily than duration. It produces a single number per employee, usually over a rolling 12 months, which organisations use to flag absence patterns for review.

The formula is:

B = S² × D

where S is the number of separate spells of absence in the period, and D is the total number of days absent.

What the squaring does

The arithmetic is the entire point. Take four employees who each missed exactly ten days in a year:

• One absence of 10 days → 1² × 10 = 10
• Two absences totalling 10 days → 2² × 10 = 40
• Five absences totalling 10 days → 5² × 10 = 250
• Ten separate single days → 10² × 10 = 1,000

Same ten days, scores a hundred times apart. For comparison, a single uninterrupted absence lasting a full working year scores around 240 — less than five two-day absences.

The reasoning is operational rather than moral: one planned two-week absence can be covered, while ten unpredictable single days disrupt scheduling ten times over. Whether that justifies a hundredfold difference in how an employee is treated is the argument the formula has generated for forty years.

Read more: A Complete Rundown On The Bradford Factor

Origin and thresholds

Its origin is murkier than most sources admit. The name points to research associated with Bradford University School of Management in the 1980s, but the university has never confirmed it, and an alternative account credits a pharmaceutical company whose managers attended a seminar there. Treat it as a widely adopted heuristic, not a validated instrument.

There is no standard trigger threshold. Published figures — 50, 125, 200, 400 — are conventions copied between employers, not findings. A threshold borrowed from a blog will be wrong for your workforce: an organisation with high part-time or shift coverage generates entirely different distributions from an office-based one. If you use trigger points, derive them from your own absence data and review them annually.

Where it creates legal risk

This is the part most explanations skip, and it is the reason the formula needs careful handling.

Disability-related absence is often short and frequent — the exact pattern the formula punishes hardest. Epilepsy, asthma, migraine, Crohn’s disease, and autoimmune conditions with unpredictable flare-ups generate multiple short spells rather than one long one. A mechanical Bradford score will escalate those employees first.

In the United States, the rule is explicit. Under 29 CFR 825.220(c), employers may not use the taking of FMLA leave as a negative factor in employment actions, and the regulation states directly that “nor can FMLA leave be counted under no fault attendance policies.” A Bradford Factor scheme is a no-fault attendance policy. FMLA-protected absence must be excluded from the score, not merely considered later. The ADA adds a parallel obligation: disability-related absence may require accommodation rather than discipline, and an inflexible points system is a familiar route to a failure-to-accommodate claim. State and city sick leave laws generally bar retaliation for protected use as well.

In Canada, the duty to accommodate to the point of undue hardship applies to absence connected to disability, and human rights tribunals look unfavourably on automatic thresholds applied without individual assessment.

In the United Kingdom, where the formula is most widely used, disability-related absence engages the duty to make reasonable adjustments — commonly by raising or disapplying the trigger for that employee — and pregnancy-related absence must be disregarded entirely.

It also encourages presenteeism. An employee two points from a warning comes to work unwell, spreading illness and lowering output. Unions have made this criticism for years, and it carries more weight after several years of public health messaging that said the opposite.

Using it responsibly

• Exclude protected absence before scoring, not after — FMLA, statutory sick leave, pregnancy-related absence, jury duty, bereavement and any absence under an agreed accommodation
• Treat the score as a prompt for a conversation, never as an automatic sanction. The formula identifies a pattern; a human has to find out why
• Never automate the outcome. A system that issues warnings at a threshold will, sooner or later, issue one to someone it shouldn’t
• Set thresholds from your own data and publish them, so nobody is scored against a rule they couldn’t see
• Review the distribution, not just the outliers. If a department’s scores cluster high, the problem probably isn’t individual
• Record the absence type, since you cannot exclude what you never categorised