Anna Totterdell
Projects Director
This is not a theoretical article. It is an account of what happened when we worked with a mid-market professional services business to fix a single broken process.
The details have been anonymised. The numbers have not.
The business
A professional services company. Around 120 employees, turning over approximately £18 million. They run an ERP system, a separate CRM, a finance platform, and a project management tool. The business is healthy, growing, and - by any external measure - well managed.
But their monthly financial close was killing them.
What we found
The finance team's monthly close took between three and four working days. Every month. Without fail. And most of that time was not spent on analysis or decision-making. It was spent on reconciliation - making numbers from different systems agree with each other.
Here is what the process actually looked like:
Revenue was recorded in the CRM when a deal was closed. But the CRM's definition of "closed" did not match the ERP's definition of "delivered." And the finance platform's definition of "recognised" was different again. So every month, someone in the finance team exported data from all three systems, loaded it into a master spreadsheet, and spent the better part of two days manually reconciling the differences.
Project costs sat in the project management tool. But not all costs were captured there - some were entered directly into the finance system by the accounts team, and some were estimated in the ERP based on standard rates that had not been updated in eighteen months. The second day of the close was largely spent identifying and resolving discrepancies between these three sources.
On day three, the finance director reviewed the assembled numbers, flagged queries, and sent them back to the team for investigation. Many of these queries were the same ones that had come up the previous month, because the underlying causes had never been fixed.
The output of this three-to-four-day process was a board pack that leadership received a week after month end. By the time decisions were made based on those numbers, they were already stale.
What we did
We did not replace any systems. We did not introduce new software. We worked with what they had.
First, we mapped the actual data flows - not the intended ones. We traced how revenue, cost, and margin data moved between the four systems and identified every point where manual intervention was required. There were fourteen separate manual steps in the close process. Eleven of them existed because the systems were not exchanging data reliably.
Second, we normalised the definitions. "Closed" in the CRM was aligned with "delivered" in the ERP and "recognised" in the finance platform. This was not a technical change - it was an IT and process strategy decision that had been deferred for years. It required the commercial director, the operations lead, and the finance director to sit in a room and agree on what each term meant. That conversation took ninety minutes.
Third, we built the data and systems integration layer. Revenue data now flows automatically from CRM to ERP to finance, with consistent definitions and timestamps at each stage. Cost data is captured once and propagated. The reconciliation spreadsheet - the one that consumed two days every month - is no longer needed, because the numbers agree by design.
Fourth, we created a live reporting dashboard. The board pack data that previously took three days to assemble is now available in real time. Discrepancies are flagged automatically as they occur, not discovered weeks later during the close.
The numbers
Before:
Monthly close duration: 3–4 working days.
Manual reconciliation steps: 14.
Time from month end to board pack: 7–8 working days.
Recurring data discrepancies per month: 15–25 (many repeated from previous months).
Finance team hours consumed by close: approximately 60 hours per month across three people.
After:
Monthly close duration: same day.
Manual reconciliation steps: 3 (exception handling only).
Time from month end to board pack: 1 working day.
Recurring data discrepancies per month: 2–3 (flagged and resolved automatically).
Finance team hours consumed by close: approximately 8 hours per month.
That is a reduction of over 50 hours per month in the finance team alone. Across a year, that is more than 600 hours returned to a three-person team - time they now spend on forward-looking analysis instead of backward-looking reconciliation.
What nobody expected
The measurable outcomes were clear. But the less quantifiable changes were equally significant.
The finance director stopped dreading the first week of every month. The board started receiving information they could act on while it was still relevant. The commercial and operations teams stopped arguing about whose numbers were right, because there was now one set of numbers - visible to everyone, in real time.
And the project delivered something else: proof. Proof that the approach works. Proof that you can improve operations meaningfully without replacing systems. Proof that weeks of targeted work can remove years of accumulated friction.
Within two months of the close process going live, the operations director asked if we could do the same thing for their resource allocation workflow. The finance director asked about business automation for the VAT return preparation. The commercial director wanted to connect the pipeline data to the capacity model.
None of those conversations would have happened if the first project had not delivered visible, measurable results on a process that everyone understood and everyone had been frustrated by.
What this means
This is one process in one company. The specifics will be different in yours. But the pattern is consistent:
A process that costs more time than anyone has quantified. Manual steps that exist because systems were never properly connected. Data that disagrees across platforms because definitions were never aligned. A team spending days on reconciliation instead of analysis.
The fix is not dramatic. It is targeted, specific, and measurable. And the results speak for themselves.
