1,570 Person-Days a Year Back: How a Manufacturing Group Rebuilt Finance Reporting After an Acquisition
This case study shows how a global manufacturing group took back an estimated 1,570 person-days of manual finance work a year, about 3.6 times the cost of the delivery in its first year. After a major acquisition it replaced five allocation processes, a manual SAP BI and BPC to Excel to Power BI relay and regional forecasting spreadsheets with one governed finance backbone on its Databricks platform. Month-end moves from weeks to days, and certified Power BI models replace shadow reporting.
Executive Summary
A global manufacturing group takes back an estimated 1,570 person-days of manual finance work a year, about 3.6 times the cost of the delivery in its first year, after moving month-end from weeks to days and five allocation processes into one on a governed finance backbone built on the Databricks platform it already runs. A five-week analytical phase produced the blueprint, and a 20-week fixed-price delivery built the finance data products, closed by a two-week UAT run in parallel with the old process and two weeks of hypercare. A team of seven, about five FTE in total, delivered both phases.
Client: a global manufacturing group
Service line: Financial Analytics
Duration: a five-week analytical phase, then a 20-week fixed-price delivery
Team: seven specialists, about 5 FTE: a data engineer, a delivery leader and two BI consultants full time, a data solution architect at 0.5 FTE, and an infrastructure engineer and a data advisor at 0.25 FTE each
The Situation: Five Allocation Processes After an Acquisition
After a major acquisition, the manufacturing group ran two finance models side by side, with different cost structures, different profitability models and multiple ERP instances. Around 15 people across several continents fed the month-end by hand, some for two days a month and some for the whole month. Allocation logic lived in five disconnected processes across legacy systems, regional files and local templates, and SG&A reporting depended on Access databases and manual reconciliation. The legacy reporting platform had been built by people who had since left, so much of its KPI and allocation logic could only be understood by reverse engineering it. Users exported raw extracts and rebuilt dashboards in Excel, each region kept its own version of the same report, and differences were closed with manual adjustment lines every month.
What We Did: One Finance Backbone on Databricks
One backbone for every report and forecast. The planned reporting projects drew on the same sources and the same transformation logic, and building them separately would have duplicated an estimated 65 to 80% of the effort. The manufacturing group's reports and forecasts now run on one governed foundation.
Built on the platform the group already runs. The finance backbone sits on the manufacturing group's Databricks platform on Azure: dbt transformations governed in Unity Catalog, loads through Azure Data Factory, and certified Power BI semantic models on top. We build the finance data products on that platform, alongside the team that runs it.
Month-end automated at the source. Direct ingestion from SAP ECC and automated GR55 onboarding replace manual extracts and offline templates. FX comes from one Bloomberg-based rate table, daily and at month-end, intercompany eliminations run as rules during ingestion, and daily snapshots trace every variance to its movement.
One definition of performance. A single model holds the KPI logic, P&L structures, hierarchies, ownership and access. Customer, product and geography hierarchies are harmonised across the group, and the allocation and KPI rules are rebuilt as version-controlled dbt transformations with named owners.
Two reporting layers and governed Excel. Executives get comparable, export-ready views, operational teams get drill-down analytics by product, customer and region, and finance teams pivot in Excel on the same certified Power BI semantic models with the KPI logic locked.
Forecasting on the same foundation. Regional templates and Power Query consolidations give way to one governed forecasting pipeline.
How We Worked: An Analytical Phase, Then Fixed-Price Delivery
The engagement opened with a five-week analytical phase that produced a decision-ready blueprint with sequenced delivery options and risk-weighted trade-offs. Delivery followed at a fixed price over 20 weeks, in two-week sprints, with acceptance criteria for every deliverable. UAT ran for two weeks as a parallel run of the new process against the old Access-based one, two weeks of hypercare followed, and the client's Product Owner signed off each component in Azure DevOps.
The Result: 1,570 Person-Days a Year Back, 3.6 Times the Cost in Year One
An estimated 1,570 person-days of manual finance work taken back every year. That is about 3.6 times the cost of the delivery in its first year, freed from extracts, Access databases, allocations and reconciliations for analysis.
Month-end from weeks to days. Automated GR55 onboarding and direct SAP ECC ingestion shorten the close and remove its failure points.
Five allocation processes consolidated into one, governed and owned, so month-end runs on one process across the group.
More than 80% of data objects shared across reporting domains, starting from the management reporting layer, and certified executive and operational models replace the regional copies and shadow Excel reports.
The Honest Lesson: Recover the Logic Before Governing It
The allocation and KPI rules had to be recovered before they could be governed: the people who built the legacy platform had left, so the rules were reconstructed by reverse engineering and from what individual team members knew. After a merger, the largest data risk tends to sit in the workarounds people have normalised, which makes governance the core deliverable.
%402x.svg)
