
Latest episode
We draw the close as it runs handoff by handoff, and remove the steps the data model can carry on its own.
Finance process optimisation reads how a group's close, forecast and reporting cycle run in practice, then removes the steps that patch the data model. Constancia, an EPM consultancy, does it for multi-entity groups whose close has stretched past the date the board pack needs.
We time each cycle over 1 full month before we redraw anything, and the timings go to the CFO first.
Each entity finishes its ledger to its own calendar, and we time every step from the last posting to the loaded trial balance.
We follow the trial balances through translation, matching and elimination, then note every point where a person re-keys a figure.
The forecast is collected from the entities each month, and we record how many workbooks it passes through before it reaches the model.
Every figure in the pack is traced to the cell it came from, and the figures typed in by hand are counted.
We list the reconciliations run each month, name who runs each one and mark which of them exist because 2 systems hold the same number.
Every step that the process document leaves out is a workaround, and we name each one with the person who performs it.
A cycle that runs differently in 2 entities is drawn twice, and both drawings go into the findings.
Our consultants have timed the close on 58 implementations over 35 years, and we draw each calendar as run before opening the process document. The document describes the close the team designed, and the calendar shows the close it runs.
Where 8 entities close in 12 working days, days 6 to 9 usually go on waiting for 1 entity and re-keying 2 others.
A workflow redesign moves those 4 days to a different desk, and a settled mapping for the 3 entities removes them.
The CFO agrees the boundary before the first cycle is timed, and it stays where it was agreed.
We redraw each cycle after the data model has been settled, and a step survives when it still has a reason. Each surviving step keeps its owner and each removed step gets a retirement date.
A redrawn close changes who does what on which day, and the decision about roles and grades belongs to the CFO and HR. We hand over the calendar as run and the calendar as designed, and the CFO decides the structure.
The 2 routes start from the same close calendar and part on the first day 2 systems disagree.
| Redesigned as a workflow | Rebuilt on the data model | |
|---|---|---|
| The step where 2 systems disagree | Moves to a different desk with a new deadline. | Disappears, because the mapping has settled the disagreement. |
| The workbook at the end of the close | Gets a named owner and a checklist. | Is retired on a date, and the model carries what it held. |
| The re-keyed figures in the board pack | Are checked by a second person. | Load from the consolidation, so the second check has nothing to find. |
| The close calendar | Shorter on paper by the days the redesign took out. | Shorter in practice, because the waiting days went with the handoffs. |
| Where the next process change begins | With another workshop when the workaround returns. | In the data model, where the definitions have owners. |
The close run to a calendar on the platform, with the routine reconciliations running themselves.
Read about the closeThe data and systems half of a wider programme, restarted from the model when it has stalled.
Read about restartsThe entity by entity read of the ledgers, mappings and workbooks that a process redraw depends on.
Read about the data assessmentEvery Sunday, 6am
Our weekly newsletter on AI in finance.
Yours to enjoy with your Sunday morning coffee.
The people who would do the work read every message and reply inside 2 working days.
Free guide
Get our AI in finance predictions for 2027
Enter your details below and we will send a free copy to your inbox right now.
Check your inbox for your copy of the predictions.
We use cookies to improve the site and measure our adverts.Read the cookie notice
Strictly necessary
Keeps the site working and remembers these choices. Always on.
Statistics
Counts visits and the pages read, so we can improve the site. Never used for adverts.
Advertising
Lets Google and LinkedIn measure our adverts and show them to people who have visited.