A spreadsheet earns its place honestly. It is fast to build, free to change, and understood by almost everyone who has to use it. The problem is not that spreadsheets are bad tools — it is that the moment they stop being enough rarely announces itself. Nobody schedules a meeting to declare that the sheet has failed. It just gets slower to trust, one small workaround at a time, until the business is running on something nobody would design on purpose.
The version-conflict problem
The clearest early sign is that more than one copy of the truth exists at the same time. A file gets emailed, edited, emailed back, and now there are two versions with different numbers in the same cell. Someone works from a local copy while someone else works from the shared one. The workaround — naming files "final," "final_v2," "final_v2_actually" — is not a filing habit, it is a symptom. A spreadsheet works when there is exactly one current version and everyone is looking at it. Once that stops being true, the sheet still runs, but it has already stopped being reliable.
Manual reconciliation is a hidden tax
The second sign is quieter and more expensive: time spent every week comparing one sheet against another, or against a system, to make sure they agree. This reconciliation work rarely shows up as a line item anywhere — it is absorbed into someone's Tuesday afternoon and never counted as a cost. But it is a cost, and it scales badly. The busier the business gets, the more there is to reconcile, at exactly the moment there is less spare time to do it properly.
One person becomes the system
The third sign is the most dangerous because it looks like competence rather than a problem. Over time, one person accumulates the formulas, the macros, the tribal knowledge of which tab means what and why a certain cell must never be touched. The spreadsheet works, but only because that person is available to interpret it. When they are on leave, or leave the company, the business discovers that the system was never really the spreadsheet — it was that person's memory of the spreadsheet.
The signs worth acting on
None of these signs alone means it is time to replace the sheet immediately. Together, and repeating, they mean the real cost has moved from the spreadsheet itself to the workarounds built around it. Worth watching for specifically: reconciliation that takes longer each quarter, a formula nobody currently on the team can explain, and a decision that got delayed because the person who understands the file was unreachable.
What replaces it does not have to be custom
Recognising the signs is the easy part; deciding what comes next is where businesses often overcorrect, jumping straight to a fully custom build when a configured, off-the-shelf platform would remove the same friction for a fraction of the cost and time. That choice — and the tradeoffs each path carries — is worth working through properly rather than guessing at under deadline pressure, which is exactly what our guide to buying, building or configuring software sets out to do. Our software team can also look at the specific spreadsheet in question and say plainly whether it has actually failed, or just needs one workaround removed.
