Spreadsheet Statistics 2026: Usage, Errors & What They Cost

Spreadsheets run the world's businesses — and quietly break them. Here's a sourced roundup of how many people use spreadsheets, how often they contain errors, and what getting it wrong actually costs.

Spreadsheets are the most-used business software on earth and among the least-audited. The same tool that runs the quarterly forecast also ships silent mistakes into board decks and bank wires. Below are the numbers worth knowing in 2026 — each with its source — on how widely spreadsheets are used, how often they're wrong, and the price of the errors.

How many people use spreadsheets

Exact counts are hard because Microsoft doesn't break Excel out of the Office/Microsoft 365 suite, and Google doesn't publish standalone Sheets numbers. But the scale is not in doubt:

• Excel: widely estimated at 750 million to over 1 billion users worldwide. Microsoft has publicly referenced a figure around a billion Office users, and analysts commonly place Excel in the high hundreds of millions. Treat any single number as an estimate.
• All spreadsheets: counting Excel, Google Sheets, Apple Numbers, LibreOffice and others, the working estimate is roughly 5 billion spreadsheet users globally.
• In business: spreadsheets remain the default analysis tool for a majority of companies, used across finance, ops, sales and reporting long after "real" BI tools are bought.

The takeaway isn't the precise headcount — it's that spreadsheets are the substrate almost every business decision passes through at some point.

How often spreadsheets are wrong

This is where the numbers get uncomfortable. Multiple independent bodies of research, spanning decades, converge on the same conclusion: most consequential spreadsheets contain errors.

• 94% of business spreadsheets contain errors. A 2024 review led by Prof. Pak-Lok Poon, published in Frontiers of Computer Science, analysed spreadsheet-quality studies spanning ~35 years and found the overwhelming majority of business-critical spreadsheets carried errors.
• ~88% carry errors in the long-running research of Prof. Raymond Panko (University of Hawaii), whose audits of operational spreadsheets became the field's most-cited benchmark.
• ~5% cell error rate. Studies put the per-cell mistake rate around 5%, consistent with the general 2–5% error rate humans hit on complex, repetitive cognitive tasks — which compounds fast across thousands of cells.
• ~50% of models used by mid-sized and large businesses contain defects material enough to change the result.

The pattern behind every one of these numbers is the same: spreadsheets let anyone build a load-bearing calculation with no tests, no review, and no record of what changed. The error rate isn't a people problem — it's a tooling problem.

What spreadsheet errors cost

Most spreadsheet errors are invisible. The ones that aren't have moved markets and made headlines:

• Citigroup — $900 million (2020). A payments error routed through a spreadsheet-driven interface sent roughly $900M of its own money to Revlon's lenders instead of a small interest payment; a court initially let the recipients keep much of it. One of the most expensive "fat-finger" mistakes on record.
• JPMorgan — the "London Whale" (2013). The bank's own internal review found that a risk model run in Excel — with values copied by hand between sheets — understated risk, a factor in trading losses that exceeded $6 billion.
• Fidelity Magellan — $2.6 billion (1994). A missing minus sign in a spreadsheet flipped a large net capital loss into a gain, leading the fund to project a dividend it then had to retract.

These are the visible tail of a much larger, quieter cost: mispriced quotes, wrong headcounts, double-counted revenue, and decisions made on a number nobody re-checked. Across an economy running on ~5 billion spreadsheets, the aggregate runs to billions a year.

Why it keeps happening — and what changes it

Spreadsheets fail for structural reasons, not careless ones: formulas break silently when rows shift, copy-paste detaches a number from its source, and there's rarely a second set of eyes before a figure reaches a decision. The fix isn't "be more careful." It's removing the fragile step.

That's the shift behind asking your data questions in plain English instead of hand-building formulas. When a tool like Quiriz turns "what was Q4 revenue by region?" into a real query it runs against your actual data — and shows the result — there's no VLOOKUP to misalign and no cell reference to drift. You can read more on why teams get different numbers from the same file and using AI to analyse data in Excel.

Sources

• Poon, P.-L. et al. (2024), "A survey of spreadsheet quality," Frontiers of Computer Science — reported by Phys.org (94% of business spreadsheets contain errors).
• Panko, R., University of Hawaii — spreadsheet error research (~88%); see the spreadsheet-research literature at Tuck School of Business (Dartmouth).
• Salesforce / Forbes (2014), "Sorry, Your Spreadsheet Has Errors (Almost 90% Do)".
• Citigroup / Revlon (2020–2021): U.S. federal court ruling on the $900M wire error, widely reported (Reuters, Bloomberg).
• JPMorgan "London Whale" (2013): the bank's internal task-force report and the U.S. Senate Permanent Subcommittee report reference the Excel-based risk model.
• Fidelity Magellan (1994): contemporaneous reporting on the missing-minus-sign dividend retraction.

Figures are drawn from published research and reporting as of August 2026. Population estimates for spreadsheet and Excel users vary by source because vendors do not publish standalone counts; ranges are given where estimates differ.