Let’s start with a defense: Excel is great software. For calculations, analysis, and anything one person works on alone, there is hardly anything better, and we regularly advise against replacing spreadsheets that work with expensive applications. The problem is not Excel. The problem is that spreadsheets grow quietly. The little calculation helper from 2019 has become a tool that orders, deadlines, and payroll depend on, and nobody noticed the moment it happened. There are signs, though, by which that moment can be recognized.

Sign 1: Several people work in the same file

Excel is built for one editor. As soon as three people work in the same file, the familiar contortions begin. The file sits on the network drive and is “currently locked”. Copies circulate with names like Planning_final_v3_NEW. Someone overwrites yesterday’s change, and nobody notices until the customer calls.

The online variants with shared editing only solve the locking problem. What is missing stays missing: who is allowed to change what? Who changed what, and when, and why? A spreadsheet has no per-column permissions and no per-record history. With two users that is bearable; with eight it becomes a permanent state of follow-up questions and repairs.

Sign 2: Only one person still understands the formulas

Every grown spreadsheet has its archaeology: hidden helper columns, a VLOOKUP onto a sheet called “old_do_not_delete”, a formula with seven nested IFs. As long as the person who built it is around, everything works. Their vacation is a risk, their resignation an emergency.

Strictly speaking that is not an Excel problem but a documentation problem. Excel makes it especially delicate, though, because business logic and data live in the same file, and one accidentally overwritten cell can destroy both. A single wrong sort over a partial range, and rows no longer line up, with no error message anywhere.

Sign 3: Money or deadlines depend on the spreadsheet

The critical question is: what happens if a value in this spreadsheet is wrong? If the answer is “we spend an hour recalculating”, everything is fine. If the answer is “a customer gets a wrong invoice” or “we miss a deadline”, the spreadsheet carries a responsibility it was never built for. And errors are the norm, not the exception. Ray Panko (University of Hawaii) reviewed field audits of operational spreadsheets from 1995 to 2004: 94 percent of the 88 spreadsheets examined contained at least one error.

A spreadsheet validates nothing on its own. It does not warn you when a delivery date is earlier than the order date, when a customer number exists twice, or when a required value is missing. People compensate with attention, and attention runs short precisely on stressful days, which is exactly when mistakes are most expensive.

Sign 4: Data gets copied in and copied back out

If the spreadsheet receives its data by copy and paste from the inbox and hands its results over by copy and paste into the invoicing tool, it is no longer a tool but a way station. Each of those manual steps costs time and invites errors, and the spreadsheet itself goes stale between transfers.

This pattern is a double signal. It argues for replacing the spreadsheet, and it shows what the replacement should look like: an application that talks to the neighboring systems directly, instead of merely being a prettier spreadsheet.

What “replacing” actually means, and what it costs

The replacement for a grown spreadsheet is rarely a big system. Usually it is a small web application: a database instead of rows, input forms with validation instead of free cells, role-based permissions, a change history. The spreadsheet itself is the best specification you could ask for, because the requirements are already written down, in columns and formulas.

Realistically, such an application costs between €8,000 and €25,000, depending on how much logic lives in the spreadsheet and how many systems get connected, plus running costs. That is far more than an Excel license, and that is exactly why the replacement only pays off when several of these signs apply at once. Replacing a spreadsheet one person uses twice a month is a waste of money.

When you should stay with Excel

Stay with the spreadsheet if one person uses it alone, if a mistake only costs rework, and if no data is regularly copied in or out. Stay with it, too, while the process is still changing rapidly: a spreadsheet can be restructured in five minutes, an application cannot. Excel is the best tool there is for prototyping a process. Only once a process is stable, and several people and real money depend on it, is it ready for its own software.

A pragmatic middle step we often recommend: keep the spreadsheet, but automate its most dangerous manual steps, such as the import from the inbox or the export into the invoicing tool. That costs a fraction and defuses the biggest sources of error.