An attendance policy is the written document setting out what an employer expects regarding presence at work, how absence and lateness are reported, and what happens when expectations aren’t met.

It’s often confused with absence management, which is the broader practice — the measuring, the conversations, the support. The policy is the written rule set that practice is applied through. One is a document; the other is a discipline.

What a policy is actually for

Not deterrence. An attendance policy doesn’t stop people from being ill, and treating it as a deterrent produces presenteeism rather than attendance.

Its real job is predictability: making the response to a given situation knowable in advance, by everyone, before anyone needs it. Two employees in different departments with the same facts should get the same outcome. A manager facing an awkward situation at 8 am on a Monday should know what to do without improvising.

That’s the test to hold any attendance policy against. Not “is it strict enough” but “does it produce the same answer twice.”

What it needs to contain

• The reporting procedure — who to contact, by when, and how. “As soon as possible” is not a rule; “by 9 am to your direct manager, by phone or message” is
• Definitions — what counts as an absence, a late arrival, an early finish, a no-show. These differ, and a policy that treats them identically will be applied unevenly
• Evidence requirements — what’s needed, after how many days, and who pays for it
• An explicit protected-leave carve-out — statutory sick leave, FMLA and equivalents, accommodation-related absence, jury duty, bereavement. Stated plainly, not implied
• Triggers — what level of absence prompts a review, and that a review is a conversation rather than a sanction
• The escalation path — the stages, who conducts each, and what evidence is considered
• Appeal rights and who hears them
• Who can exercise discretion, on what grounds, and who records it

That last point is the one most policies omit. Discretion is necessary, but undocumented discretion is indistinguishable from inconsistency when someone challenges it later.

Read more: How to Write an Employee Attendance Policy?

No-fault and point systems

A no-fault attendance policy assigns points to each occurrence regardless of reason, with defined consequences at thresholds. A late arrival might score half a point, an unexcused absence one, a no-show two, with a warning at four and dismissal at eight.

The appeal is obvious: it removes manager judgement, so it looks fair and applies evenly. The hazard is the same feature. A system that ignores reason will, by design, penalise absence the law protects.

In the US this is not theoretical. FMLA leave can’t be counted under a no-fault attendance policy — the regulation says so directly. And the EEOC pursued Verizon over an inflexible no-fault policy that made no exception for disability-related absence; the company settled for $20 million, then the largest disability discrimination settlement in a single EEOC suit, and the consent decree required it to revise the policy to accommodate disability and to train the people administering it.

A point system can work. It has to exclude protected absence before points are assigned, and leave a documented route to accommodation.

What makes a policy fail

Vagueness. “Excessive absenteeism may result in disciplinary action” commits to nothing and justifies anything. It gives no notice to the employee and no protection to the employer.

Inconsistency. This is the serious one, and it’s counterintuitive: a policy applied inconsistently is worse than no policy at all. With no written standard, you argue the facts. With a written standard you’ve departed from, you’ve documented the gap between what you said you’d do and what you did — and that gap is the case against you.

No protected-leave exclusion. If the policy doesn’t say which absence is carved out, managers will count it, because the policy told them to.

Dormancy. A policy ignored for three years and then enforced against one person is the clearest evidence of selective treatment available.

No review cycle. Statutory leave entitlements change frequently, so a policy written five years ago is almost certainly wrong somewhere.

Writing one

• Write it to be applied, not to be read. Decision points, not principles
• Name the protected categories explicitly, and update that list when the law moves
• Make triggers prompt a conversation, never an automatic outcome
• Train managers on the conversation, since they’re the ones who actually run it
• Apply it to everyone, including people it’s awkward to apply it to — that’s precisely where it gets tested
• Record what was decided and why, especially where discretion was used
• Review it annually against current law and your own absence data

Frequently asked questions

What should an attendance policy include? Reporting procedure, definitions, evidence requirements, an explicit protected-leave carve-out, triggers, the escalation path, appeal rights, and who may exercise discretion.

Are no-fault attendance policies legal? In principle, but not when applied to protected absence. FMLA leave cannot be counted in one, and an inflexible policy that doesn’t accommodate disability creates significant exposure.

Can you discipline someone for being sick? For genuine illness, generally no — and protected sick leave carries anti-retaliation rules. Patterns and failures to follow the reporting procedure are a different matter.

How many absences is too many? There’s no universal answer, and a policy that names a number without context will catch people it shouldn’t. Set triggers from your own data and treat them as review points.