I’ve always had respect for dimensions, i.e., the categorization of transactions. In my first finance job, after six rounds of interviews, I was offered the chance to lead a country’s finance operations or become the global controller. Fresh out of my finance Master’s program, I chose “global controller” as my first role, mainly due to the global title.
I had three months to prepare for a US audit, which meant consolidating data from all our countries. Unfortunately, nobody had touched our dimension setup since the previous year’s audit nor implemented any semblance of regular reporting. Missing departments, country codes, or outdated revenue splits plagued the 100,000+ transactions that I reviewed, making the data impossible to slice accurately. It took countless Red Bulls, sleepless nights, and massive batch changes to get the financial statements ready to meet the audit deadline.
Luckily, we passed the audit and then had a full year to rebuild the master data to ensure those mistakes wouldn’t happen again. The next year’s audit took a quarter of the time.
But it happened again in another role, and I missed it. Sales had decided to split their Account Executive coverage from one round-robin pipeline into three distinct funnels (EU → EU, UK&I, Partnerships) after one of our monthly forecast updates.
Cue the first monthly review: Finance reported a sales ACV increase, but Sales rejected it, claiming EU numbers had dropped. After rounds of back-and-forth, we realized we were looking at entirely different datasets, as Finance included partnership deals (which were higher ACV) while sales did not. The new pipeline split had made our standard funnel views useless and turned the cost of acquisition per channel into a manual calculation. To align the data, our team then needed to:
Update the CRM to require new dimensions and rebuild all reports
Add dimensions to the ERP to properly reflect each new division
Re-sort every current customer into the new categories
Create new pricing tables and currency conversions
Redefine key SaaS metrics per channel (e.g., ACV for each segment)
Reconfigure ad platforms to optimize algorithms for each region
Correct historical reports to match the new structure
It took a month to fix…
Two months later, priorities shifted, and we made an entirely new categorization structure.
The lesson? Any time you change the data you track, whether that is decided during a budgeting process or throughout the year, you will need to restructure your data across your tech stack to ensure you can cleanly pull your data. To do this properly, you need to:
Lean into the Dimension Discussion
Walk people through each dimension (data categorization). Yes, people’s eyes will glaze over but it’s essential that all teams understand your data splits. These definitions shape how you view your business by defining which regions, countries, departments, etc. that you track. To do this:
Define which dimensions to track. Start with the basics: country, region, department. But, do you want to track sub-departments? How about regions within countries? Each dimension you want to track adds complexity for accurately capturing each transaction, but can also provide more detailed performance drill-downs.
List out each option in each dimension. Create a spreadsheet showing all relevant combinations. You’ll quickly see how multiplying dimensions adds complexity (see image below).
Identify which systems track what data. Ensure you know which dimensions are tracked in each system.
Align the naming. Align the wording and naming conventions in every system (e.g., BCN vs. Barcelona)
Don’t overcomplicate. Resist pressure from single-person sub-departments to create their own setups (e.g., marketing splitting into demand gen, etc.). Things change quickly and you don’t want to be constantly tweaking these splits.
By putting real focus into your data structure, you will simplify your reporting process for many years.
Tip: Disallow combinations in each system that aren’t valid. For instance, if you don’t have a partnership team in the UK, prevent anyone from selecting “partnership” as a UK department in your system.
Example:
Create a Data Dictionary to Link Past, Present, and Future
A data dictionary is simply a guide to data definitions. It lists all dimensions, the dates they are active, and how you define them internally. For example, it details which countries are included in the EMEA definition and what differentiates product marketing from content marketing. Once you are ready to make your changes, you should:
Systemize your data dictionary. Build a master overview in your FP&A system or data warehouse to consolidate all valid dimensions into one overview. You will need to query this often, so make it accessible!
Define dimensions clearly. Avoid confusion by documenting exactly what each dimension contains. This is especially helpful for regional or categorical changes over time.
Map old and new metrics. Export your data from each system, align it side by side with your new dimensions, and match each to the updated metric structure.
Date-stamp changes. Include the effective date at the top of each change to enable you to generate past reports in the new format and new ones in the old format (if needed).
Having this data in one place will vastly simplify your historical and future reporting.
Tip: Integrate your SaaS unit definitions into this table as well to maintain a central repository of all data definitions.
Update Your Systems with New Fields and Reporting
As you finish the data dictionary, review all systems (including HRIS, accounting, and CRM) to identify which changes needed across your company. To do so:
Extract all dimensional data. Pull out the dimensions from each system by exporting the relevant data.
List the reports that need updating. Inevitably, many reports and filters will need to be changed in nearly every system. Prioritize and triage these reports.
Establish the data granularity. Oftentimes, you will have different data granularities depending on your systems and priorities. Make sure that each report goes down to the lowest level available and ensure that any overview reports match the right dimension level. (For example, if you only have expenses categorized as “marketing,” don’t allow the report to split them further.)
Decide how to track old dimensions. Determine how historical data will be handled under the new dimensions. Establish whether to reflect changes in your existing reports or create separate reports using the old mapping for accurate historical comparisons.
Make the changes. Have the relevant owner update the dimensions in each system.
Communicate the updates. Provide clear instructions to each team on when and how to use the new dimensions.
These steps are straightforward and will ensure that there are no gaps in your end-of-month reporting.
Tip: Make all updates on the same day to ensure consistency across systems.
Test Reporting and Establish Ownership
Once everything is set up, test reporting to ensure both old and new dimension setups work as expected. This is where you will identify any outliers or mismapping. To do so:
Run a financial report using the old mapping. Ensure these numbers match previously released figures. Review all your key metrics across dimensions to confirm consistency.
Run past and present financials using the new mapping. Expect some changes and identify and explain any differences.
Run your reforecast with the new dimensions. Your reforecast should use the same data dictionary. Run all your reforecasts with the old and new dimensions to ensure your planning aligns with your actuals.
Test reporting at different levels of granularity. All reports should reconcile on the highest levels of categories (say region) even if the sub-categories differ or are absent. Your reports should be able to handle differing levels of granularity seamlessly.
Include changes in your board deck. Add an appendix to explain how key metrics or board reporting are impacted. Keep it concise but provide enough detail for reference.
Establish clear ownership of master data. Have the finance team take responsibility for maintaining and updating the data moving forward.
Run these reports and reviews as soon as you change your dimension setup to avoid reconciliation issues and inaccuracies later.
Tip: Create an automated report that flags any missing dimensional data (or any old mappings) and have it regularly reviewed. This will save significant time during later reconciliations.
In Conclusion
In fast-moving companies, you must balance locking down your data structure with the need for flexibility. Solidifying your data structure through dimensions enables easier reporting and clear alignment across teams. However, each update to your dimension setup risks muddying your historical financials and, as you can see, involves a significant amount of work in each part of your tech stack and enuing data reconciliations. So,make changes infrequently and execute them thoroughly to minimize confusion.
