Almost every company we walk into has one. A spreadsheet holding the real administration, usually maintained by one person, usually with colours only that person understands. People point at it sheepishly when we ask.
There is no need for that. That spreadsheet is proof somebody solved a problem without waiting for anyone else. That is exactly the behaviour you want.
Where it does go wrong
It becomes a problem the moment more than one version exists. Someone emails it on, someone else adds to it, and two weeks later nobody knows which one is right. That is where the mistakes come from, not from Excel itself.
The second problem is that the knowledge lives in one head. That person goes on holiday and the process stops. That person leaves and it is gone. We have helped clients who spent weeks reconstructing how their own pricing agreements worked after a departure.
The third is that a spreadsheet stops nothing. One wrong sort, one formula that shifts a row, and you find out when the client calls.
What we do with it
The first thing we do is read the spreadsheet. Not to replace it, but because it holds more process knowledge than any document. The exceptions are in there. The real order of steps is in there. Often there is even a tab recording what once went wrong.
Then we build software around that process, with the same freedom and without the risk:
- One place where it is right, with permissions per person.
- A history: who changed what, and when.
- Checks in the places where it went wrong before.
- Exporting to Excel still works, because for analysis it is still the best tool there is.
When to just leave it alone
Sometimes the answer is to keep the spreadsheet. If one person uses it, if the process changes every quarter and if no money rides on it: leave it. Building software around it costs more than it returns, and we will say so.
The rule of thumb we use: as soon as two people want to be in it at once, or as soon as a mistake costs money, it is time. Not before.
