Spreadsheets are wonderful and most businesses should use more of them. They're the fastest way to work something out, and a good one has often saved somebody hiring software they didn't need.
The trouble starts at the point where one stops being a tool and becomes infrastructure. That's a different job, and a spreadsheet was never built for it.
The eight signs
One of these on its own is normal. Three or four together is the pattern.
- Two people have it open and somebody's edits vanish. A spreadsheet was designed for one person at a time. Shared editing helps and it doesn't fix the underlying assumption.
- There's a copy called final-v3-USE-THIS. The moment there are two versions, nobody can say which is true, and somebody is making decisions from the wrong one.
- Somebody rebuilds the same report every month. The work isn't the numbers, it's the assembling. That's a job software does and a person shouldn't.
- A tab is hidden because it breaks when touched. One person understands it. Your business now depends on them not leaving.
- You email it round to find out who has the latest. There is no latest, and everyone is working from a different picture.
- It takes ten seconds to open. Size alone is the mildest sign here, and it's usually the first one people notice.
- Everyone can see everything. Including salaries, margins and customer details, because a spreadsheet has no idea who's looking.
- One wrong paste went unnoticed for weeks. Nothing in a spreadsheet checks anything. A wrong number looks exactly like a right one.
None of these are about arithmetic. A spreadsheet is still better at that than anything you'd build.
What they're really telling you
The same four problems sit underneath all of them.
- More than one person needs it at once. This is the big one and it's what spreadsheets fundamentally aren't.
- Nobody can see who changed what. No history, so a mistake can't be traced and an argument can't be settled.
- Not everyone should see everything. Permissions don't exist.
- Nothing validates anything. No rule stops a date going in the wrong column, so errors are silent by default.
Those four are the definition of the job having outgrown the tool. It isn't a criticism of spreadsheets, any more than you'd blame a van for not being a warehouse.
Try these before building anything
Genuinely worth doing first, and often enough that's all it takes.
- One copy, in shared storage, with everything else deleted. Not archived. Deleted. Half of the problems on that list are two versions.
- Turn on validation. Restrict the columns that keep going wrong to a list or a date. It takes ten minutes and stops the silent errors.
- Lock the formula cells. Most spreadsheet damage is somebody typing over a formula without noticing.
- Check whether software you already pay for does this. Job tracking, stock and scheduling are often already in a product you own and never switched on.
- Look for an off-the-shelf product. If your process is ordinary, buying beats building comfortably.
When it's genuinely time to build
Three conditions, and you want at least two of them.
It's load-bearing. The business stops if the file is wrong or missing. At that point it's infrastructure sitting in a format with no backup discipline and no history.
Several people need it at once, with different views. A shared spreadsheet with hidden columns is a permissions system held together with hope.
It costs real hours every week. Count them honestly. Four people at twenty minutes a day is about 340 hours a year, and that's the number to hold a build price against.
What replacing it looks like
Smaller than people expect, and it shouldn't be a platform.
The usual shape is one screen that holds what the spreadsheet held, with the four missing pieces added: several people at once, a record of who changed what, permissions, and rules that stop bad data going in.
Custom software here starts at $5,000 with a fixed price agreed before the build begins, and this kind of build is often at the lower end. Larger or more custom projects are quoted separately.
Keep the spreadsheet afterwards. It's still the right tool for working something out before you decide it's permanent, and the good habit is doing exactly that and then moving it once it proves itself.