Planned and actual: the reconciliation that shows you where the hours go

Echipa HR 365 · reviewed 2026-09-02 · 5 min read

Reconciliation means comparing, for each person and each day, what was planned with what happened, and classifying every difference. Differences are not errors — they are information. What matters is that none of them reaches the final timesheet without somebody having seen and labelled it.

Why one of them alone is not enough

Companies that plan shifts but never check delivery are, in effect, paying the plan — including shifts nobody covered. Those that record only attendance, with no plan, can never say whether coverage was sufficient, because there is nothing to compare against. The two together answer the useful question: was the right person in the right place?

The five types of difference

What can appear and what each means
The differenceWhat happenedWho resolves itAn alarm signal if…
Planned, not deliveredThe shift was not coveredManagerIt recurs on the same weekday
Delivered, not plannedSomebody came who was not in the planManagerIt happens often — the plan is not being used
Longer durationThey stayed beyond scheduleManager, then HRIt is constant for the same person or post
Shorter durationThey left earlyManagerThere is no recorded reason
An unrecorded swapTwo people swapped shifts between themselvesManagerIt surfaces at reconciliation rather than before

The monthly procedure

Five steps, in the same order every month

  1. System or HR — generates the list of differences, not the whole timesheet — you work only on what does not match
  2. Team manager — classifies each difference against the five types, within two days of month end
  3. Manager — adds the reason where the type alone does not explain it (left early — why)
  4. HR — checks only the unclassified differences and the recurring ones, not all of them
  5. HR — closes the month; after closing, any correction leaves a trace and a reason

The first step is what makes the procedure sustainable. In an ordinary month, differences are under 10% of the cells. A manager who receives the list of differences finishes in twenty minutes; one who receives the whole timesheet and has to read it never reads it.

What you read from the pattern of differences

Closing the month and why the written trace matters

Once closed, the month is not changed silently. Any later correction stays visible, with the author, the date and the reason. It is not a measure of distrust: it is the only way that, six months on, anybody can reconstruct why a day looks different from the report sent at the time.

What you do when the differences are many and constant

Above roughly 15% of cells differing, the problem is no longer execution but planning. A plan that constantly departs from reality does not describe the work, and reconciliation becomes a monthly transcription of what happened anyway. In that situation, the right move is to rebuild the plan from the actuals, not to insist that actuals move towards the plan.

The practical check takes ten minutes: take the last three months and count how many times the same position had a difference on the same weekday. If Friday evening is always undercovered, the plan tells a story nobody believes any more, and people have adjusted on their own. Their adjustment is usually closer to the truth than the plan.

Who sees the reconciliation and who does not

The result of reconciliation is operational information, not a disciplinary file. The manager sees their team, HR sees the aggregate and the recurring differences, the employee sees their own days. Nobody sees a list of colleagues with deviations — the moment such a table circulates is the moment people start covering for each other, and the data becomes useless.

What reconciliation is not

It is not a tool for controlling people and it produces no sanctions. Presented that way, people start reporting exactly what was planned, whatever happened — and you lose the only source of information about what the work actually looks like.

The test that shows which direction the process has taken at your place: look at the number of differences reported over the last three months. If it drops steadily towards zero, it is unlikely you fixed the planning; far more likely you taught the team that differences bring uncomfortable questions, and they have disappeared from the report rather than from reality.

Do I reconcile monthly or weekly?

Weekly for classification, monthly for closing. A week later the manager remembers what happened; a month later they reconstruct from assumptions.

What if the manager does not classify the differences?

Do not classify them yourself, by assumption. Unclassified differences stay marked as such in the closed timesheet — visible in the report. It is the only mechanism that produces an effect, because the problem becomes theirs rather than yours.

Planned-versus-actual reconciliation, generated automatically on the differences

The manager receives only the cells that do not match, classifies them in one click, and the monthly timesheet closes with a reason on every deviation.

See the timesheet module

Free account, every module for 7 days, no card required.