The spreadsheet that quietly runs your organisation

You know the one. It has forty tabs, a colour code nobody documented, and a column called DO NOT DELETE. It was built four years ago as a temporary fix, and it is now the most load-bearing piece of software in the organisation.

Nobody planned this. It happened one reasonable decision at a time.

How it happens

The pattern is always the same, and it is never anybody's fault:

1. The real system cannot do one specific thing.
2. Someone sensible builds a small spreadsheet to cover the gap.
3. It works, so it grows a second use. Then a third.
4. Other people start depending on its output.
5. It is now infrastructure, and it has no owner, no tests, and no backup.

Each of those steps is defensible. The end state is not.

The three costs nobody counts

The obvious cost is time: the hours spent maintaining it. That is the one people quote, and it is the smallest of the three.

The shape of a coordination week

Where the week goesShare

1. 1Re-keying data that already exists elsewherehigh
2. 2Chasing the person who holds the missing fieldhigh
3. 3Reconciling two versions of the same truthmedium
4. 4Answering “which of these numbers is right?”medium
5. 5The work the team was actually hired to dowhat is left

The ordering is the point. Almost none of that is the job, and none of it shows up on a timesheet as a cost, because every line of it looks like diligence.

The re-keying cost. Data arrives in one system and gets typed into the spreadsheet by hand. Every keystroke is a chance to be wrong, and the two copies begin drifting apart the moment they exist.

The trust cost. When two systems disagree, work stops while someone finds out which is right. Multiply that by every decision that touches the number.

The key-person cost. One person understands the formulas. When they are on leave, a part of the organisation runs on hope. When they resign, it runs on archaeology.

Why "just buy software" usually fails

The instinct is to replace the spreadsheet with a proper tool. Sometimes that is right. Often it is not, because the spreadsheet is not the problem. It is the symptom. It exists because a real process has a gap in it, and if you buy software without closing that gap, one of two things happens.

Either the gap reappears as a new spreadsheet next to the new software, or the team quietly keeps using the old one and you are now paying for both.

The spreadsheet is a scar. It tells you exactly where the organisation was cut. Replacing it without asking what cut it is how you end up with the same wound and a subscription.

What to do instead

Start by finding out what the spreadsheet is actually compensating for. Follow the data in and the data out, and ask why each hop exists.

Usually you find one of three things: a system that cannot export, a system that cannot import, or two teams who were never given a shared definition of the same word. Only the first two are software problems, and the third is the most common.

Fix the process gap first. Then, and only then, decide whether anything needs building.