The hidden cost of spreadsheet dependency

Let us begin by being unfair to our own argument: we like spreadsheets. They are the most flexible business tool ever created, they cost almost nothing, and a well-built one is a thing of genuine craft. We build them for clients regularly. What we object to is not spreadsheets — it is spreadsheet dependency: the state a business drifts into when a temporary worksheet becomes permanent operational infrastructure. The two look identical from the inside, and the difference only shows up when it starts costing money.

How dependency happens

No business decides to depend on spreadsheets. It happens one reasonable decision at a time: a stock list here, a job tracker there, a workaround for a system limitation, a report someone built for a one-off need that never quite ended. Three years later there are nine workbooks, several maintained by people who have since left, some with formulas that broke silently, and a month-end ritual that merges them all by hand. Each step along the way was sensible. The destination is fragile, opaque and expensive.

The four costs nobody has line-itemed

• The reconstruction cost. Every month, someone re-assembles the truth from fragments. Count the hours, then multiply by twelve, then by their wage — and then count the senior hours spent checking numbers that should simply be right.
• The error cost. Broken formulas fail silently; overwritten cells have no undo history; two versions of the same figure circulate until a decision is made on the wrong one. Spreadsheet errors are rarely dramatic — they are small, frequent and compounding.
• The key-person cost. The workbook's logic lives in its author's understanding. When that person is unavailable — holiday, illness, resignation — the business does not lose a file; it loses the process.
• The opportunity cost. The quietest cost of all: decisions made late or on instinct because the real numbers were days away. Few businesses ever put a price on a late buying decision or a mispriced job, but that is where the largest losses hide.

The four warning signs we look for

1. A workbook has a name people know by heart. "The stock master" is no longer a tool; it is infrastructure.
2. Only one person fully understands a sheet. Ask who can rebuild it from scratch. Silence is the warning.
3. Data is re-typed into it from other systems. Double entry is where errors and hours both breed.
4. Decisions wait for the spreadsheet. When the tool sets the decision tempo, the tail is wagging the dog.

Our position: evolve, don't condemn

Here is where we differ from the technology vendors who sponsor most writing on this topic: the answer is rarely "buy a platform" and never "abandon spreadsheets". The answer is to evolve the spreadsheet's role. Make the system of record something structured and reliable, and let spreadsheets do what they are brilliant at — analysis, one-off calculations, working views. In practical terms: consolidate the fragmented workbooks into one governed source with agreed definitions, automate the feeds that currently rely on re-typing, and keep the flexible layer for the thinking work. In most businesses we assess, a surprising share of the pain disappears from that reorganisation alone, without any new software at all.

The place to start is not a purchase; it is an audit. Which workbooks are load-bearing, who understands them, and where do they disagree with the accounting system? That exercise takes days, not months, and it tells you whether you have a tools problem or a foundation problem. We run it as a fixed-price engagement, and we are rarely the ones who end up surprised by its findings.

What the audit actually delivers

Because "audit" can sound abstract, here is what ours produces in practice. An estate map: every workbook in the business listed, with its purpose, its owner, its last-modified date and the systems it feeds from or feeds into — the first time most owners have seen the whole picture on one page. A load-bearing classification: each workbook marked as infrastructure (the business stops without it), operational (used weekly, replaceable) or historical (archive it), which is what turns a vague anxiety into a worklist. A reconciliation check: the figures in the critical workbooks compared against the accounting system and, where relevant, physical counts — with every mismatch listed and traced. And a recommendation set in three tiers: what to retire this month, what to consolidate this quarter, and what genuinely justifies a system decision. The whole exercise is deliberately sized so that even if you act on none of it immediately, you will never again be guessing about a foundation your business stands on.