Excel as a shadow ERP: how to spot it and what it costs

The symptom

Somewhere in your business there is a file. It is called something like Planning_2026_final_v3_NEW.xlsx, it lives on a network drive, and when the colleague who maintains it goes on holiday, part of the company grinds to a halt.

That file holds which order has which priority. Which machine is booked when. Which customer gets which special terms. Which batch is still sitting in quarantine. None of it lives in the ERP you pay a five-figure maintenance fee for every year — or it does live there but is no longer correct, because the truth migrated into the spreadsheet long ago.

That is a shadow ERP: a spreadsheet that has quietly become the actual system of record. Nobody decided this. It just happened.

The mechanism

Shadow ERPs are not born of laziness; they come from a genuine gap. The official system can't do something — a particular report, a piece of planning logic, a field that simply doesn't exist. Someone builds the matching spreadsheet in five minutes. At first, that is exactly right and pragmatic.

The tipping point comes when other people start to rely on that spreadsheet. A personal aid becomes a shared source. Because a spreadsheet is more flexible than any ERP, it grows: a column here, a second tab there, a formula that chains three cells together. Two years on, a small information system has taken shape — only without any of the things that make a system a system.

Because Excel has no permissions model worthy of the name. No real history of who changed what and when. No validation to stop someone typing "see email" into a column of numbers. No safe simultaneous editing by several people — instead you get _copyMueller and _copySchmidt, which someone has to merge back together in the evening. And no integration: whatever the spreadsheet holds has to be maintained twice, because the ERP beside it knows nothing about it.

How do you recognise it? When a file's name carries final and a version number. When a process only keeps moving as long as one particular person is present. When the same figure lives in two places and regularly shows two different values. When nobody can answer "what is the reliable source here?" on the spot.

The cost

The price of a shadow ERP is so high precisely because it never shows up on an invoice. It spreads across three silent line items.

Double maintenance and rework. Every piece of information that has to sit in both the spreadsheet and the ERP gets touched twice. Take an employee who spends two hours a day reconciling data between the sheet and the system and resolving the contradictions. Two hours × 220 working days × 45 euros = 19,800 euros a year for work that simply wouldn't exist with one clean source. And that is one person; usually it is several.

Key-person risk. The spreadsheet is fully understood inside exactly one head. When that person is out — holiday, illness, resignation — the knowledge is gone. What that costs in a pinch depends on what hangs off the sheet. If it holds the production schedule, a single day of absence can cost more than the 19,800 euros of annual rework.

No traceability. When a figure is wrong, Excel can't reconstruct who changed it, when, or why. In the harmless case that is an annoyance. For price-relevant or batch-controlled data it is a compliance problem — and expensive if it ever comes to a dispute with a customer or an auditor.

The fix

The wrong response is to ban the spreadsheet. It exists because it fills a real gap — take it away without closing the gap and you sabotage the operation.

The right path starts with a sober inventory: which shadow spreadsheets exist, which decision depends on each one, and which gap in the official system does it fill? Only once that is on the table do you decide, spreadsheet by spreadsheet and deliberately.

For many, it is enough to add the missing capability in the system of record — a custom field, a report, a small workflow. For others, a lightweight intermediate solution with a proper database, permissions, and history is the better route, cleanly connected to the ERP so that each figure exists only once. The common denominator: a single reliable source per piece of information, instead of two that contradict each other.

Which spreadsheet needs which treatment is exactly the sorting work that stands at the start of a process analysis — making the real data flow visible before anyone touches the software.

The next step

If you honestly can't say what the reliable source for an important figure in your business is, the first step is not new software. It is an honest map of the file that actually runs your company.