When a Spreadsheet Is Still the Right Answer
There is a certain kind of advice that treats spreadsheets as a problem to be solved. Any process running in one is assumed to be a candidate for a proper system, and moving off them is treated as progress by default.
That is often right and it is not always right. Plenty of businesses have replaced a working spreadsheet with a system that is slower, more expensive and less understood, and called it a modernisation.
What spreadsheets are genuinely good at
It is worth being honest about why they are everywhere. They are not a failure of planning. They are extremely good at several things.
- Changing shape quickly. A new column takes seconds. In most systems it takes a change request.
- Being understood. Nearly everyone can read one, and many people can fix one.
- Costing nothing extra. No licence, no implementation, no vendor conversation.
- Handling the genuinely one off. The analysis you will do once, the model for a decision you are making this month.
For work that is low volume, single owner, and still being figured out, a spreadsheet is frequently the correct engineering decision.
The four questions that decide it
Rather than a rule about spreadsheets, ask four questions about the process.
How many people touch it. One is fine. Four editing the same file is where it starts to hurt.
What happens if it is wrong. A planning model that informs a conversation is different from the file that decides what customers get invoiced.
Is it the authoritative record. If the real answer to who our customers are lives in a spreadsheet, every system built around it is reporting on a copy.
Is it stable. A process that changes every month is one a spreadsheet handles better than a configured system.
The question is not whether it is a spreadsheet. It is how many people depend on it, and what happens when it is wrong.
The signs it has outgrown itself
Some symptoms are reliable. Several versions in circulation with dates in the filename. A monthly ritual of reconciling it against another source. One person who is the only one who can fix it. Formulas nobody dares touch. A copy emailed around because the shared version keeps getting locked. Errors that were found by a customer rather than internally.
Any two of those together is usually the point at which the cost of keeping it exceeds the cost of moving.
Make the spreadsheet better before you replace it
Quite often the process is fine and the file is simply badly built. Before funding a system, a few hours of tidying can buy you another two years.
Separate the data from the calculations and the presentation. Use one row per record, with no merged cells. Use data validation so the status column has five values rather than eleven. Protect the formula cells. Put the version, owner and last review date at the top. Write a short note explaining what it is for and how to use it.
None of that is glamorous and all of it reduces the error rate immediately.
If you do move, move the process, not the file
The most common migration mistake is recreating the spreadsheet inside a new system, complete with its workarounds, and then wondering why nothing improved.
Use the move as the moment to ask what the process should be. Which steps exist only because the spreadsheet made them necessary. Which fields nobody has filled in for a year. Which of the eleven statuses actually mean something.
Standardise first, then build. Otherwise you have paid to make the old shape permanent.
Keep the ones that earn their place
Even after a system arrives, some spreadsheets should survive, and that is not a failure of the project.
Modelling and scenario work. One off analysis. The working file somebody uses to think. The thing that is genuinely easier this way.
What matters is that they are not the authoritative record for anything, that they do not quietly become a second system, and that anyone relying on one knows it is a working file rather than the truth.
The practical position
Treat spreadsheets as a legitimate tool with a known set of weaknesses, rather than as either a scandal or a permanent answer. Review the important ones once a year and ask whether the four questions still give the same answer.
That review, and the honest conversation about which processes have outgrown their file, is a common piece of work for us. Sometimes the recommendation is a system. Fairly often it is a better spreadsheet and a written process, which costs a fraction as much and solves the actual problem.
Not sure whether a process has outgrown its file? Book a consultation, or visit www.expertechsolution.com.
