The same scene repeats every month-end. The sales lead sends a spreadsheet with the month’s revenue. The warehouse manager sends another with stock levels. The accountant calculates a third number that matches neither. The first half hour of the meeting goes to arguing whose number is right, instead of deciding what to do about it.
This isn’t anyone being careless. It’s the inevitable result of how small businesses start working with data: every department spins up its own spreadsheet because that’s the fastest way to get their job done. The problem only shows up once those spreadsheets get compared to each other.
Why five files never agree
Three causes show up in almost every business I’ve worked with on this problem.
- Manual copy-paste. A number moves from one system into a spreadsheet by copy-paste. Every manual step is a place where a mistake, an omission or a wrong version can creep in.
- Many people, many copies. A file called “sales_final_v3.xlsx” gets emailed around, someone makes a copy, edits it and sends it back under a different name. Soon nobody knows which version is the newest.
- No one owns the number. When the revenue figure lives in three different files with three different people, nobody ends up responsible for it being correct. Everyone trusts their own sheet and suspects everyone else’s.
Add to that the fact that spreadsheets are usually built at different times with different filters (one includes VAT and another doesn’t, one counts the order date and another the shipping date), and you end up in a situation where every number is “correct” but none of them answer the same question.
What a real fix actually looks like
The usual reaction is to build a sixth, even more thorough spreadsheet that rolls up all the others. That fixes the problem temporarily and recreates it three months later, because the root cause is still there: the data still lives in several different places and someone still has to stitch it together by hand.
The real fix is simpler than it sounds: one place where the number actually lives, and everything else that displays that number (reports, spreadsheets, dashboards) reads from it instead of keeping its own copy. In practice that means a proper database or a small data warehouse that the sales system, the warehouse system and accounting feed automatically. Excel doesn’t disappear, it just becomes a window onto the data instead of the data’s home. Nobody copies numbers into a spreadsheet by hand anymore; the spreadsheet pulls them automatically from one source.
It’s the same difference as a shared, live document versus copies emailed back and forth. When everyone reads the same document, there’s no question about which version is correct.
How to start without it turning into a year-long project
The biggest mistake businesses make is trying to fix everything at once: every data source, every department, one big system. That’s expensive, slow, and usually stalls out. The better path is to start small.
- Find the number that causes the most arguments. Usually it’s revenue, stock levels, or some other monthly figure that keeps getting disputed in meetings.
- Map where that number actually comes from. How many systems and spreadsheets feed into it, who updates them, and when.
- Build a pipeline for just that one number. Connect the original source (sales system, online store, accounting software) automatically to one place that this number now comes from.
- Let the old and new number run side by side for a while. Once they match, trust is established and you can move on to the next number.
- Repeat with the next painful number. Each subsequent one is faster to add, because the core infrastructure already exists.
Three months in, you have at least one number everyone trusts, instead of waiting until year-end for a big system that may or may not ever get finished.
In short
Five spreadsheets that don’t agree isn’t a technical problem, it’s an ownership problem: nobody is responsible for a single truth. The fix isn’t a sixth spreadsheet, it’s one place everything else reads from. Build it number by number, starting with whichever one hurts the most.
If there’s a number in your business that gets argued over every month, see what my data engineering and BI work covers, or write to me and describe which number that is for you. We’ll look together at the simplest way to get it into one place.