Spreadsheets Aren't the Enemy. Ungoverned Ones Are.

There is a genre of business writing in which spreadsheets are the villain. Spreadsheet hell, spreadsheet sprawl, the spreadsheet that nearly bankrupted a company. I want to push back on this, because it sends the wrong lesson. Spreadsheets are not the problem. Ungoverned spreadsheets acting as systems of record are the problem, and the distinction matters because the fix is not "stop using spreadsheets."

What spreadsheets are genuinely good at

Spreadsheets are the best thinking tool most businesses have. Ad-hoc analysis: "what would margin look like if we dropped this product line?" What-if modeling: change one cell, see the ripple. One-off reports for a meeting next week. Prototyping a calculation before you commit to building it properly. And crucially, they let a small team do all of this without filing a ticket or waiting for a developer.

That last point is why they spread. They are not a failure of discipline, they are a success of accessibility. A company with zero spreadsheets is usually a company where analysis waits in a queue, which is its own kind of expensive.

Where they stop being good

The line is simple: spreadsheets for thinking, databases for remembering.

A spreadsheet is a bad place for anything the business depends on being true. The failure modes are specific and familiar. Version forks: final_v3_REAL, final_v4_boss-edits, and nobody can say which one produced the number in the board deck. No audit trail: a figure changed last Tuesday and the file does not know who did it or why. Single-machine memory: the company's pricing logic lives in a workbook on one laptop, and that person is on leave. Silent breakage: someone deletes a column, a formula quietly shifts, and the error surfaces a quarter later in an audit.

None of those failure modes are spreadsheet flaws. They are what happens when a thinking tool gets promoted to a system of record without anyone deciding to promote it.

Governance without a platform

At 20 to 150 people you do not need a data platform. You need three rules applied to the handful of spreadsheets that actually matter, which you can count on one hand.

One named owner per critical sheet. Not a team, a person. Their job is to know if it breaks and to say when it is stale.

Inputs separated from outputs. The raw numbers come from a synced source, on a schedule, into a tab nobody edits. The thinking happens in formulas and views on top. The moment someone re-types a raw number by hand, the sheet has become a rumor.

Critical numbers get pulled, not typed. If the sheet needs revenue, it reads revenue from where revenue lives. If the source is not available on a schedule, that is the actual project, and the spreadsheet is just the symptom.

Everything else, the ad-hoc analysis, the what-ifs, the meeting prep: let people spreadsheet freely. That is the tool working as designed.

The uncomfortable summary

If your business would notice a spreadsheet going dark for a week, that spreadsheet is a system of record and it needs the boring protections of one: an owner, a source, a place where history is kept. If nobody would notice, it is a thinking tool and it should stay free.

The companies that get this right are not spreadsheet-free. They are spreadsheet-governed, and the difference shows up exactly once a year, when the numbers in the deck match the numbers in the bank.