Your spreadsheet runs the business. That's the risk.

No one set out to run a business on a spreadsheet. It started as the fastest way to get the truth working — a pricing model, a production schedule, a pipeline tracker, a set of supplier terms — built by someone who understood the work better than the tools did. That is not a mistake. Excel can be an excellent engine.

The problem begins when the spreadsheet quietly becomes infrastructure. The day it stops being a convenient model and starts being the system of record — the one everyone else reads, keys into, and depends on — it is production software without production controls.

That is the risk worth taking seriously, because it is almost never urgent until it suddenly is.

The failure is never the spreadsheet itself

It is what the spreadsheet cannot do, once it matters:

• It has no single version of the truth. Two people keep two versions; a third keeps a working copy they have not told anyone about. The "correct" schedule is whichever file was open most recently, and nobody can prove it.
• It is not auditable. A formula changes, a row is overwritten, a number is pasted over another. If there is a trail, it lives in someone's memory. When an audit, a dispute, or a handoff turns on what was true last Tuesday, there is no answer.
• It does not scale with people. The productivity gains that made the spreadsheet feel great for one domain expert become friction for ten people who each own part of the spreadsheet and none of it.
• It is invisible to the rest of the operation. The precious model dunks data in and out by hand, so the system next to it never hears what the spreadsheet knows — and every handoff re-enters, re-checks, and re-drifts.

These do not crash loudly. They compound quietly, until a regulatory requirement, an audit finding, or a scaling event forces a rebuild under pressure — at the worst possible time, with the least possible budget.

How to tell whether it is infrastructure yet

Run this check on your most important spreadsheet. Ask:

• If two people opened it right now, would both be looking at the truth? (Versions and untracked working copies mean no.)
• Who is the single person who can explain every formula in it? (If it is one person's head, it is memory-held work — see our earlier note on where work depends on memory.)
• If it vanished tonight, what breaks tomorrow morning? (If the answer is "the ordering," "the schedule," or "payments," it is infrastructure.)

If any of those feel uncomfortable, the spreadsheet has crossed from tool to system of record. It is not Excel's fault. The tool outgrew its design.

The useful change is usually small

The instinct is to rip it out and build something grand. Resist that. In the operations we have run firsthand, the high-value move is almost always to wrap the spreadsheet with structure rather than replace the thinking inside it:

• Keep the spreadsheet as the model if the logic in it is good — most of the time it is, and throwing it away throws away real accumulated judgment.
• Give it a controlled life. One canonical version, explicit ownership, a promotion path from "working copy" to "the one people can trust."
• Connect it so the systems it already feeds stop importing by hand, and vice versa.
• Add the basic controls it lacks — who can change it, what changed, and when—so it becomes auditable without losing the speed that made it valuable.

The spreadsheet should not disappear overnight. It should be reinforced in place: same familiar workflow, but with a single truth, a trail, and a guardrail. That is a change a small team can feel and verify — and it does not require a development project or new infrastructure.

Where it is genuinely fine

Not every spreadsheet needs rescuing. A working document for one person, a scratch model, a one-off analysis — leave those alone. Automating a spreadsheet that does not matter yet is exactly the kind of premature work that adds moving parts for no return. The risk is specifically the spreadsheet that has already become production, whether anyone has admitted it or not.

One honest boundary, so this does not read as a pitch: if your spreadsheet is not yet infrastructure — if losing it would not stop anything tomorrow — then you are probably fine for now. Re-run the check when the spreadsheet starts deciding things. That is the moment to act, not earlier, and not later.

If a spreadsheet in your operation has crossed into infrastructure and you want a second pair of eyes on how to reinforce it, talk it through with us or write to hello@ridgepathsystems.com. We will tell you honestly whether it needs work, and what we would do first.