Too Small to Count

No timekeeping system records everything. Some amount of time is always disregarded, by rounding, by a grace period, by a threshold below which nothing is counted, or simply by nobody looking.

The practical lesson in “Too Small to Count” is that a timesheet only becomes reliable through a process people can operate consistently. For teams exploring hourly timesheet template, this workforce platform can add time and project context, provided collection is proportionate, access is limited and every consequential inference receives human review.

That is reasonable and unavoidable. What is not reasonable is the way the threshold is usually set: by what is administratively convenient, applied in one direction, and never multiplied out.

For a separate benchmark relevant to “Too Small to Count”, consult the FASB accounting standards. Use it to test record quality, approvals, retention, employee rights and exception handling against the real workflow rather than treating a software report as self-explanatory evidence.

Where the disregarding happens

Rounding at the punch, to the nearest five, six or fifteen minutes.

Grace periods, where a punch within a window of the scheduled start counts as on time — which is fine for attendance and becomes a problem when it also determines pay.

Minimum increments, below which a period of work does not register at all.

And the informal version, which is the largest: activity that happens and is never put into any system because everybody treats it as too small to bother with.

The two tests that matter

Is it symmetric? A threshold that disregards up to seven minutes in both directions is neutral over time and defensible. One that rounds the start forward and the end backward takes up to fourteen minutes a day from every shift and is indefensible in most jurisdictions regardless of how small each instance is.

Check your own configuration rather than your policy. The asymmetric version is a default in several products and it is frequently in place without anybody having chosen it.

Does it survive multiplication? Seven minutes twice a day, across four hundred people, is twenty-four thousand hours a year. The claim that an amount is too small to count is a claim about the instance, and liability and cost are both properties of the total.

What the legal position tends to be

It differs by jurisdiction and the direction is consistent: small amounts of working time may be disregarded only where recording them is genuinely impractical, and the disregarding must not systematically favour the employer.

The "genuinely impractical" part has largely collapsed. It made sense when a supervisor was writing times on a card. With a system that records to the second, the argument that a regular, predictable, measurable period cannot be captured is difficult to sustain, and tribunals and inspectorates have said so.

A regular period — the same seven minutes every day, for the same reason — is the hardest to defend, because its regularity is itself proof that it could have been measured.

Setting a defensible threshold

State it, rather than inheriting it. Say what is disregarded, how much, in which direction, and why.

Make it symmetric and make it small. Round at the latest point rather than at every punch, because rounding each of four daily punches compounds and rounding the daily total does not.

Review it against the multiplication, annually, with the figures above. A threshold whose annual total is several full-time posts is not a threshold; it is a pay policy that nobody agreed.

Who is allowed to set it

A threshold determines pay, so it is a term of employment rather than a system setting, and in many places it cannot be changed unilaterally once established.

In practice it is frequently altered by whoever administers the time system, during an upgrade, with no record of the decision. That is the version that produces an unexplained change in paid hours across a site and a round of grievances nobody can account for. Any change to rounding, grace periods or minimum increments should be a documented decision with a named owner and a date, and the old setting should be recorded alongside the new one.

The honest alternative

For the regular, predictable edge — the walk, the change, the queue — the clean answer is not a threshold at all. It is to include the time in the paid day and shorten the nominal shift to match, which costs the same as paying it and removes the question permanently.

Organisations resist this because it looks like conceding something. What it actually concedes is a liability they were already carrying, in exchange for a record that matches reality and a number nobody can recalculate against them later.

Where the threshold is hidden

Most organisations have never set one and have several anyway, distributed across the configuration: a grace period on arrival, a tolerance on the schedule match, a rounding increment, a minimum overtime block.

Each was set separately, none was set as a policy about what is too small to count, and together they are the organisation's actual position on the question. Listing them in one place is usually the first time anybody sees the aggregate, and the aggregate is frequently considerably larger than any threshold the same people would have agreed to set deliberately.