The Flat Week

Sort any organisation's timesheet data by pattern and the largest single group will be five eight-hour days. In a workforce doing varied work in varied conditions, this is the one shape the week almost certainly did not have.

The record described in “The Flat Week” should be created close enough to the work that people are not reconstructing a polished week from memory. When teams assess further details are available here for fireable offenses, they should keep entry, project selection and correction simple, while explaining which optional activity data is collected and how employees can review it.

It is not fraud and it is rarely even dishonesty. It is what a person submits when they are reconstructing a week they did not record, under a deadline, against a contract that says thirty-seven and a half hours. The flat week is the honest output of an unrecorded week, and it tells you about the capture process rather than about the person.

For a separate benchmark relevant to “The Flat Week”, consult the Project Management Institute learning resources. 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.

What the flatness hides

Variation in both directions. The eleven-hour Wednesday when a site visit overran and the five-hour Friday when the work ran out. Those cancel in the weekly total and vanish in the flat submission, and what is lost with them is every operational signal the data could have carried.

Allocation detail is lost too. A flat week usually arrives with flat allocation: the whole week on one code, because splitting it by memory is harder than splitting it by hours. If the organisation bills, claims or capitalises by code, the flat week is quietly wrong in the place where wrongness is most expensive, and it is wrong in a way that no total-level check will ever catch.

Why it is particularly dangerous in claims

In a payroll-only context the flat week costs little. The person is salaried, the hours are nominal, and the record is a formality nobody will ever read.

The moment the same record supports a grant claim, a client invoice or a capitalisation judgement, the flatness becomes a finding. Funders look for exactly this pattern because it is the signature of reconstruction, and a run of identical weeks is treated as evidence that the records are not contemporaneous. The consequence is not usually a dispute about the hours; it is disallowance of the claim on record-keeping grounds, which is a worse outcome because it does not matter whether the hours were right.

Detecting it without accusing anybody

Compute, per person per period, the variance of daily hours, and the number of weeks in which the daily pattern is identical to the previous week. Both are two lines of code against any export.

Report them as a population, not as a list of names. What you are looking for is proportion: if sixty percent of weeks are perfectly flat, you have a capture problem affecting the whole organisation, and naming individuals for an organisational failure produces defensiveness and no improvement. If three people out of two hundred are flat and everybody else is not, that is a different conversation and it belongs with their manager rather than in a report.

The causes, in order of how often they are the cause

Capture happens at the end of the week rather than during it, which is nearly always the main one. The interface takes too long to use more than once a week — eight clicks per entry means nobody makes five entries. Nothing prompts during the week. The person has no way to record on a phone from a site. Rounding rules are coarse enough that fifteen-minute variation disappears anyway. And occasionally the work genuinely is flat, which happens in fixed-shift environments and is identifiable because the clock data agrees.

Fixing the capture, not the submission

Nothing in a flat week is improved by an instruction to be more accurate. The person did not record the detail; they cannot now produce it, and asking them to will generate a more varied fiction rather than a true record.

What works is cutting the cost of recording at the time. One-tap start and stop from a phone. Prefilling from the rota or the calendar so the person edits rather than types. A midweek prompt — Wednesday afternoon, not Friday — which halves reconstruction distance on its own. And reducing the number of codes, because a short list can be used from memory while a long one cannot. The flat week is a symptom of distance between the work and the record, and every one of these closes that distance by a day or two, which is all it takes.