What is spreadsheet governance?

Spreadsheet governance is the practice of validating, recording and measuring the data inside the spreadsheets a critical business process depends on — without changing the spreadsheet itself. It is not a tool category by itself; it’s a set of controls applied to a model that already exists and that nobody is going to rebuild.

Why this needs a name at all

Most companies don’t have a “spreadsheet problem.” They have a finance team, a quality lab, a planning function, each with one or two spreadsheets so central to how the work gets done that the business would struggle to operate without them for a day. Those models were never meant to be temporary — they became permanent because they worked, and nobody scheduled time to replace something that wasn’t broken.

The gap is that “working” and “governed” are different things. A spreadsheet can produce the right answer every time for years and still have no record of who entered the last change, no check that stops an out-of-range value, and no way to prove — to an auditor, a regulator, or your own leadership — that the number on the screen is the number it should be. Spreadsheet governance is the name for closing that gap without touching the first part.

What it actually consists of

Four things, consistently:

1. Validation at entry. A field has a valid range, format or list of acceptable values, defined once and enforced every time someone types into it. A dosage can’t be ten times too high. A percentage can’t exceed 100. The check happens before the bad value is saved, not after someone notices it downstream.
2. A recorded history. Every change is logged automatically — who, what, when, and what the value was before. Not a version-numbered file copy, a field-level record that survives the file being resaved a thousand times.
3. Controlled periods. Once a period, a batch, or a submission is marked final, its values are locked. A correction after that point is a new, recorded change, not a silent edit to history.
4. A measure of reliability. Not just whether a value was within range, but how often it needed correcting, and whether that’s getting better or worse over time, field by field.

None of this requires moving the spreadsheet anywhere or asking the person who built it to change how they work. It requires a layer that sits between the spreadsheet and wherever its data ends up — see how IT-TIDY works for one way to build that layer without a migration project.

Each of the four pieces answers a different question, and it’s worth being precise about which one you’re actually missing, because the fix is different for each:

• No validation shows up as errors that are only caught downstream — in a failed batch, a rejected invoice, a finding during an audit — never at the moment they were typed.
• No recorded history shows up as a question nobody can answer: “who changed this, and when?” The file has an answer somewhere, in someone’s memory or an email thread, but not in the spreadsheet itself.
• No controlled periods shows up as a number that was “final” last week and is quietly different this week, with no record that it moved or why.
• No measured reliability shows up as a team that feels like data quality is improving, or getting worse, but has no way to show it — every conversation about it is an opinion, not a number.

It applies outside finance

Spreadsheet governance gets discussed most often in finance, because consolidation and reporting spreadsheets are old, well-known pain points with regulatory attention attached. But the same four pieces apply anywhere a spreadsheet drives a decision with real consequences. A capacity model in network planning has the same need for validated inputs and a change history as a budget model — the stakes are just different (a wrong number affects a rollout instead of a forecast). A formulation sheet in a chemicals plant needs the same controlled-period logic a finance close does, just triggered by a batch release instead of a month-end. The practice is identical; only the field names and who’s affected change.

What it isn’t

Spreadsheet governance isn’t an inventory. Cataloguing every spreadsheet in a company and flagging the risky ones is a reasonable first step for an audit team, but it stops at visibility — it tells you where the risk is without reducing it. Governance is the active part: once you know a spreadsheet is critical, governance is what you actually do about it.

It also isn’t a migration plan. “We’ll govern this model better once we move it into [enterprise system]” is a sentence that gets said in a lot of planning meetings and rarely finishes with the move actually happening, because the system never fits the model as well as the spreadsheet did on day one. Governance doesn’t wait for that project. It applies to the spreadsheet that exists today.

And it isn’t the same as data quality in the abstract. “Our data should be accurate” is a goal. Governance is the mechanism: specific rules, on specific fields, checked at the moment of entry, with a record kept of what happened when a rule caught something.

A concrete example

A formulation sheet in a chemicals plant has a dosage field with an approved range of 1.5–2.5%. Without governance: someone types 4.8% by mistake, the sheet saves it, the batch runs, and the error is caught — if it’s caught — only when the output fails quality control downstream. Nobody can say who entered 4.8% or when, because the file has been resaved a dozen times since.

With governance: the 4.8% entry is rejected the moment it’s typed, with a message explaining why. If a value within range but unusual gets flagged for review instead, that review and its outcome are recorded. At the end of the month, the team can see which fields in the sheet are most often triggering corrections — not as a one-off report, but as a running measure that either improves or doesn’t.

The spreadsheet didn’t change. What changed is that the business can now answer questions about it that it couldn’t answer before.

A second example, outside manufacturing: a capacity-planning model shared by five regional teams in a telecom network-planning function. Without governance, each team keeps its own copy, edits happen by email, and the planner consolidating the group total is, in practice, choosing which version to trust. With governance, each team still edits in the spreadsheet they already use; the values are validated and consolidated into one governed database as they’re entered, and the consolidator works from one current version instead of reconciling five.

Why it’s being talked about more now

Governance failures involving ungoverned records aren’t hypothetical. EuSpRIG — the European Spreadsheet Risks Interest Group — has maintained a public catalogue of real, reported spreadsheet-error incidents since the 1990s, including cases where a single formula mistake produced a public profit warning or a multi-million-dollar budget error (EuSpRIG). On the governance side, a 2022 KPMG paper on end-user computing describes a financial-services client that had to review its EUC policy and remediate roughly 1,500 spreadsheets once a proper risk assessment was done (KPMG, “EUC proposition,” Sep 2022). Scale like that is why governance has become a named practice rather than something handled case by case.

Where to start

Governance doesn’t have to start everywhere at once. The usual sequence is: pick the one process where a wrong number would cost the most — financially, operationally, or in an audit — define what a valid entry looks like for its most important fields, and measure what happens for 30 days before deciding whether to extend it. That’s the shape of a 30-day pilot: one process, one team, a before-and-after you can actually show someone.

If you’re not sure where your own exposure is highest, a focused spreadsheet risk check-up takes about 60 seconds and gives you a starting point.