---
title: "20 spreadsheet error and data risk statistics for 2026"
description: "Research data on spreadsheet error rates, financial costs, and why businesses are moving to purpose-built software\n"
image: "https://hercules.app/blog/spreadsheet-error-data-risk-statistics/opengraph-image"
---

# 20 spreadsheet error and data risk statistics for 2026

Research data on spreadsheet error rates, financial costs, and why businesses are moving to purpose-built software

- **Author:** [Grant Gurvis](https://hercules.app/blog/authors/grant)
- **Tag:** Company
- **Published:** 2026-11-03
- **Last updated:** 2026-11-04
- **URL:** https://hercules.app/blog/spreadsheet-error-data-risk-statistics

A 2008 research review reported errors in [94% of 88 spreadsheets](https://mba.tuck.dartmouth.edu/spreadsheet/product_pubs_files/literature.pdf) examined across seven historical field audits. The finding highlights spreadsheet risks but does not establish the error rate across businesses in 2026. These risks can justify moving critical workflows into [structured applications](https://hercules.app/docs) where teams can configure validation rules, access permissions, and record-change tracking.

For businesses running operations on spreadsheets, the risk compounds over time. As files grow more complex, so does the likelihood of errors that affect decisions, budgets, and compliance.

## **Key takeaways**

- **Historical audits found widespread errors**: A 2008 research review reported errors in [94% of 88 spreadsheets](https://mba.tuck.dartmouth.edu/spreadsheet/product_pubs_files/literature.pdf) examined across seven field audits conducted between 1995 and 2004.
- **Financial errors can have major consequences**: JPMorgan's 2012 trading losses reached at least [$6.2 billion ](https://www.hsgac.senate.gov/subcommittees/investigations/rep/senate-investigations-subcommittee-holds-hearing-and-releases-report-on-jpmorgan-chase-whale-trades/)amid multiple risk-management failures, including spreadsheet model errors. Norway's sovereign wealth fund separately reported a benchmark-calculation error with an estimated [NOK 984 million impact](https://www.norges-bank.no/contentassets/ae154c4ad6484d04b290cb7be9639b9a/representantskapet_arsrapport_2023_eng_uu_web_v2.pdf?v=12062024095857).
- **Finance teams still rely on spreadsheets**: The 2025 AFP survey found that [96% of respondents](https://www.financialprofessionals.org/training-resources/resources/survey-research-economic-data/Details/afp-fpa-benchmarking-survey-report-technology) used spreadsheets for planning, highlighting their continued role in financial workflows.
- **Financial errors can be reduced**: Gartner reported a [75% reduction](https://www.gartner.com/en/documents/5211763) in financial errors among companies that digitized accounting operations and achieved high employee technology acceptance. This finding does not measure the effects of AI app builders.
- **Purpose-built applications offer configurable controls**: Businesses can use the [Hercules app builder](https://hercules.app/) to build database-backed applications with validation, permissions, and automated workflows. Teams must test these controls and implement record-change tracking when needed.

### **1. 94% of audited business spreadsheets contained errors**

Raymond Panko compiled seven field audits conducted from 1995 through 2004. A 2008 research review reported a weighted historical finding of [94% of 88 spreadsheets](https://mba.tuck.dartmouth.edu/spreadsheet/product_pubs_files/literature.pdf) containing errors. This figure covers spreadsheets used in real business operations, not test files. The audits suggest widespread errors among the spreadsheets examined, although their selection methods and error definitions varied.

### **2. 91% of KPMG-audited spreadsheets had significant errors**

A KPMG Management Consulting audit of 22 spreadsheets found that [91% contained significant errors](https://mba.tuck.dartmouth.edu/spreadsheet/product_pubs_files/literature.pdf). This 1998 study covered decision-support spreadsheets from major UK banking and financial organizations, so its findings should not be treated as a current error rate. The finding illustrates that even professionally managed spreadsheets are not immune to mistakes.

### **3. 1.79% average cell error rate across 270,722 formulas**

A 2009 study by Dartmouth researchers audited 50 operational spreadsheets and found a [1.79% cell error rate](https://mba.tuck.dartmouth.edu/spreadsheet/product_pubs_files/errors.pdf) across 270,722 formulas, including incorrect results and poor practices. When counting only errors that produced wrong results, excluding poor practices, the rate dropped to a [0.87% error rate](https://mba.tuck.dartmouth.edu/spreadsheet/product_pubs_files/errors.pdf). In a spreadsheet with thousands of formulas, even a 1% error rate can mean dozens of mistakes affecting reports or decisions.

### **4. 96% of FP&A professionals used spreadsheets for planning**

In the 2025 AFP FP&A Benchmarking Survey, [96% of respondents](https://www.financialprofessionals.org/training-resources/resources/survey-research-economic-data/Details/afp-fpa-benchmarking-survey-report-technology) used spreadsheets for planning, and [93% used spreadsheets](https://www.financialprofessionals.org/training-resources/resources/survey-research-economic-data/Details/afp-fpa-benchmarking-survey-report-technology) for reporting on a daily or weekly basis. The survey collected responses from 362 financial planning and analysis practitioners worldwide. These findings show that spreadsheets remain central to finance workflows despite growing use of dedicated planning software.

### **5. 61% of FP&A respondents identified unreliable data as a technology challenge**

The 2025 AFP survey found that [61% of respondents](https://www.financialprofessionals.org/training-resources/resources/survey-research-economic-data/Details/afp-fpa-benchmarking-survey-report-technology) identified a lack of reliable data as a challenge to using technology successfully. Problems included difficulties combining information from different systems and locations. The finding reflects respondents' reported challenges, not a measured spreadsheet error rate. It shows why businesses need dependable data and review processes.

### **6. 60% of FP&A respondents struggled with inaccessible data**

The same 2025 AFP survey found that [60% of respondents](https://www.financialprofessionals.org/training-resources/resources/survey-research-economic-data/Details/afp-fpa-benchmarking-survey-report-technology) identified inaccessible data as an obstacle to technology success. Data spread across disconnected systems can make financial planning and reporting harder. Teams may need to collect and reconcile information before using it. Structured applications can help centralize records when their integrations and access rules are configured correctly.

### **7. 71% of FP&A respondents used enterprise performance management tools quarterly**

The 2025 AFP research found that [71% of respondents](https://www.financialprofessionals.org/training-resources/resources/survey-research-economic-data/Details/afp-fpa-benchmarking-survey-report-technology) used enterprise performance management (EPM) tools at least quarterly. EPM software supports financial planning and performance reporting. However, widespread spreadsheet use continued alongside these systems. This suggests that adopting dedicated software does not automatically eliminate spreadsheet dependence. Businesses still need to address disconnected workflows and data requirements.

### **8. JPMorgan's $6.2 billion trading loss involved spreadsheet model errors**

JPMorgan Chase's London Whale trades generated losses of at least [$6.2 billion in 2012](https://www.hsgac.senate.gov/subcommittees/investigations/rep/senate-investigations-subcommittee-holds-hearing-and-releases-report-on-jpmorgan-chase-whale-trades/). Its management task force documented a [spreadsheet formula error](https://business.cch.com/srd/JPMorgan-Task-Force-Report.pdf) in the implementation of a Value at Risk (VaR) model, including a formula that divided by a sum instead of an average. The model problems contributed to inaccurate risk measurement but did not alone cause the full trading loss.

### **9. TransAlta lost US$24 million from a copy-paste error**

In 2003, TransAlta Corp. announced a [US$24 million loss](https://www.theregister.com/2003/06/19/excel_snafu_costs_firm_24m/) caused by a cut-and-paste error in Excel. The mistake affected bids submitted for electricity transmission contracts. TransAlta purchased more transmission hedging contracts at higher prices than intended. The incident demonstrates how an undetected spreadsheet mistake can create a direct financial cost during a business transaction.

### **10. Norway's sovereign wealth fund estimated a NOK 984 million benchmark error**

Norway's sovereign wealth fund reported an estimated [NOK 984 million impact](https://www.norges-bank.no/contentassets/ae154c4ad6484d04b290cb7be9639b9a/representantskapet_arsrapport_2023_eng_uu_web_v2.pdf?v=12062024095857) associated with an incorrect bond benchmark calculation. The error began in December 2021 and was discovered in February 2023. The calculation used December 1 country weights rather than the required November 1 weights. Financial reporting described the impact as approximately US$92 million. The estimate concerns benchmark performance, not a separate trading loss directly caused by Excel alone.

### **11. Fannie Mae corrected a $1.1 billion spreadsheet-related accounting error**

In October 2003, Fannie Mae corrected a [$1.1 billion understatement](https://www.washingtonpost.com/archive/business/2003/10/30/new-fannie-mae-error-rattles-market/4a0e19e1-c61a-4d2c-b6bd-7bc20a5aa98f/) of stockholders' equity after discovering computational errors in its financial reporting. Company officials attributed the mistakes to incorrect formulas in a spreadsheet used to implement new accounting rules. The correction did not change reported earnings. It was an accounting misstatement rather than a $1.1 billion cash loss, showing why financial calculations need independent review.

### **12. TIBCO's spreadsheet share-count error created a $100 million valuation difference**

In 2014, TIBCO Software disclosed a [$100 million difference](https://www.sec.gov/Archives/edgar/data/1085280/000119312514427486/d829865dex991.htm) between its initially announced equity valuation and the corrected figure in a proposed acquisition. A spreadsheet used for the transaction had counted certain restricted shares twice. The mistake affected the reported transaction valuation and prompted shareholder litigation. It illustrates how spreadsheet errors can influence major financial decisions even when individual spreadsheet entries appear correct.

### **13. 37.7% of recorded error instances involved hard-coding**

The 2009 Dartmouth study classified [37.7% of instances](https://mba.tuck.dartmouth.edu/spreadsheet/product_pubs_files/errors.pdf) as hard-coding among 483 recorded error instances. These included numbers typed directly into formulas instead of referenced from cells. This practice makes spreadsheets harder to update and audit. Researchers classified some hard-coding instances as poor practices rather than incorrect numerical results, so the percentage should not be interpreted as the share of calculations that produced wrong answers.

### **14. 32.9% of recorded error instances were reference errors**

Reference errors accounted for [32.9% of instances](https://mba.tuck.dartmouth.edu/spreadsheet/product_pubs_files/errors.pdf) recorded in the 2009 Dartmouth study. These errors occur when formulas refer to incorrect cells or ranges. A formula may still produce a number, making the mistake harder to notice. Checking references and reviewing how formulas connect across worksheets can help teams identify incorrect calculations before they affect reports.

### **15. 21.9% of recorded error instances were logic errors**

Logic errors represented [21.9% of instances](https://mba.tuck.dartmouth.edu/spreadsheet/product_pubs_files/errors.pdf) recorded in the same 2009 study. These mistakes occur when a formula uses an incorrect calculation or rule, even when its cell references are valid. They can be difficult to identify through basic formatting checks. Reviewing business rules and testing calculations against known examples can help detect these problems.

### **16. 3.4% of recorded error instances involved copy-paste mistakes**

Copy-paste errors accounted for [3.4% of instances](https://mba.tuck.dartmouth.edu/spreadsheet/product_pubs_files/errors.pdf) recorded in the Dartmouth study. These errors occurred when spreadsheet content was copied or pasted incorrectly. Although less frequent than hard-coding and reference errors in this sample, copy-paste mistakes can still affect important financial decisions. TransAlta's 2003 incident illustrates why teams should review data transfers before submitting bids or completing transactions.

### **17. 58% of recorded error instances produced wrong results**

The 2009 study classified [58% of error instances](https://mba.tuck.dartmouth.edu/spreadsheet/product_pubs_files/errors.pdf) as producing wrong results. The remaining instances involved poor practices that did not necessarily lead to incorrect calculations. Researchers also found that [48% of error cells](https://mba.tuck.dartmouth.edu/spreadsheet/product_pubs_files/errors.pdf) contained wrong results. These findings show why spreadsheet audits should distinguish between risky practices and errors that actually change reported numbers.

Applications built with structured databases can organize information using defined fields and relationships. Teams must still configure and test data validation, calculations, and access controls to reduce mistakes.

### **18. UK police service fined £750,000 for spreadsheet data breach**

In October 2024, the UK Information Commissioner's Office imposed a [£750,000 fine](https://ico.org.uk/about-the-ico/media-centre/news-and-blogs/2024/10/what-price-privacy-poor-psni-procedures-culminate-in-750k-fine) on the Police Service of Northern Ireland after a 2023 spreadsheet disclosure exposed personal information belonging to [9,483 officers and staff](https://ico.org.uk/about-the-ico/media-centre/news-and-blogs/2024/10/what-price-privacy-poor-psni-procedures-culminate-in-750k-fine). The breach occurred when a spreadsheet containing sensitive information in a hidden tab was released in response to a freedom of information request. The incident highlighted failures in checking documents before publication.

### **19. London council exposed personal data belonging to 6,528 people**

In May 2025, the UK Information Commissioner's Office reprimanded Hammersmith and Fulham Council after a spreadsheet disclosure left [6,528 people affected](https://ico.org.uk/about-the-ico/media-centre/news-and-blogs/2025/05/london-council-reprimanded-for-exposing-personal-details/), including [2,342 children](https://ico.org.uk/about-the-ico/media-centre/news-and-blogs/2025/05/london-council-reprimanded-for-exposing-personal-details/). The information had been published in October 2021 and remained exposed for almost two years. The incident involved hidden spreadsheet content that was not properly checked before disclosure.

### **20. UK Ministry of Defence spreadsheet contained hidden data on more than 18,000 people**

In July 2025, the UK Information Commissioner's Office confirmed that a Ministry of Defence spreadsheet shared in 2022 contained hidden personal data relating to [more than 18,000 people](https://ico.org.uk/about-the-ico/media-centre/news-and-blogs/2025/07/ico-statement-in-response-to-2022-mod-data-breach/). The spreadsheet concerned applicants under an Afghan relocation program. The incident was discovered in 2023. It demonstrates how hidden data can create serious security risks when spreadsheets are shared without adequate checks.

Spreadsheet access and change-tracking controls vary by product and configuration. Purpose-built applications can offer more tailored permissions, validation, and audit requirements when those features are configured correctly.

## **Moving beyond spreadsheets with Hercules**

The [Hercules app builder](https://hercules.app/) helps businesses create custom applications by describing their needs in plain language. Teams can build structured workflows with configurable controls instead of relying on disconnected spreadsheet files.

- **Data validation**: Set required fields and validation rules, then test them.
- **Role-based access**: Configure permissions and verify they work correctly.
- **Automated workflows**: Reduce repeated manual entry with tested processes.
- **Database storage**: Store structured records and, on the Business plan, schedule [daily or weekly backups](https://hercules.app/changelog/schedule-database-backups).
- **Audit requirements**: Review authentication events and implement record-change tracking separately when needed.

Hercules requires a Business or Enterprise plan for commercial use, including internal operations. AI Credits and Cloud Credits are separate, and additional usage may incur charges under the [Terms of Service](https://hercules.app/legal/terms-of-service).

Data labeling company Imjomat reports saving [2,000 team hours monthly](https://hercules.app/customers/imjomat) after replacing fragmented spreadsheet workflows. Petony Transportes built a logistics ERP that reduced repeated data entry across [three disconnected systems](https://hercules.app/customers/petony-transportes).

These are customer-reported outcomes, not guaranteed reductions in errors. Test validation and permissions before moving critical workflows into a new application.

## Frequently asked questions

### What percentage of spreadsheets contain errors?

A review of seven historical field audits found errors in [94% of 88 spreadsheets](https://mba.tuck.dartmouth.edu/spreadsheet/product_pubs_files/literature.pdf) examined between 1995 and 2004. Historical audits by KPMG in 1998 and Coopers & Lybrand in 1997 reported errors in [approximately 91% of spreadsheets](https://mba.tuck.dartmouth.edu/spreadsheet/product_pubs_files/literature.pdf) they examined. These findings do not establish a current error rate for all spreadsheets.

### How much do spreadsheet errors cost businesses?

Spreadsheet-related errors have contributed to major financial incidents. TransAlta reported a [US$24 million loss](https://www.theregister.com/2003/06/19/excel_snafu_costs_firm_24m/) in 2003. JPMorgan reported at least [$6.2 billion in trading losses](https://www.hsgac.senate.gov/subcommittees/investigations/rep/senate-investigations-subcommittee-holds-hearing-and-releases-report-on-jpmorgan-chase-whale-trades/) in 2012 amid broader risk-management failures, including spreadsheet model errors. Norway's sovereign wealth fund reported a separate benchmark-calculation error with an estimated [ NOK 984 million impact](https://www.norges-bank.no/contentassets/ae154c4ad6484d04b290cb7be9639b9a/representantskapet_arsrapport_2023_eng_uu_web_v2.pdf?v=12062024095857).

### What causes most spreadsheet errors?

In a 2009 audit of 50 operational spreadsheets, hard-coding, incorrect cell references, and formula logic accounted for the largest shares of recorded error instances. Hard-coding accounted for [37.7% of instances](https://mba.tuck.dartmouth.edu/spreadsheet/product_pubs_files/errors.pdf), reference errors for [32.9% of instances](https://mba.tuck.dartmouth.edu/spreadsheet/product_pubs_files/errors.pdf), and logic errors for [21.9% of instances](https://mba.tuck.dartmouth.edu/spreadsheet/product_pubs_files/errors.pdf). Some recorded instances were poor practices rather than errors that produced incorrect results.

### How can businesses reduce spreadsheet errors?

Businesses can reduce errors by moving critical data into structured applications with validation rules, using databases instead of disconnected files, automating data entry where possible, and implementing formal review processes. The [Hercules app builder](https://hercules.app/docs) lets teams build applications with configurable controls without writing code, although they must configure and test those controls correctly.

### When should a business replace spreadsheets with custom software?

Consider replacing spreadsheets when data is shared across teams, when errors have financial or compliance consequences, when the same data is entered in multiple places, or when spreadsheets have grown too complex to audit reliably. The [Hercules app builder](https://hercules.app/) lets businesses build custom applications for these workflows, with appropriate validation, permissions, and testing.

```json
[
  {
    "@context": "https://schema.org",
    "@type": "BlogPosting",
    "headline": "20 spreadsheet error and data risk statistics for 2026",
    "description": "Research data on spreadsheet error rates, financial costs, and why businesses are moving to purpose-built software\n",
    "image": "https://hercules.app/blog/spreadsheet-error-data-risk-statistics/opengraph-image",
    "datePublished": "2026-11-03",
    "dateModified": "2026-11-04",
    "author": [
      {
        "@type": "Person",
        "name": "Grant Gurvis",
        "givenName": "Grant",
        "familyName": "Gurvis",
        "url": "https://hercules.app/blog/authors/grant",
        "description": "Founder / CTO at Hercules",
        "image": "https://cms-media.hercules.app/57bfc56c-7f2d-4fe2-8cce-32e433d6faca/Grant%20Gurvis%20Headshot-original",
        "sameAs": [
          "https://www.linkedin.com/in/grant-gurvis/",
          "https://x.com/gurgrant"
        ],
        "knowsAbout": [
          "Engineering",
          "AI"
        ]
      }
    ],
    "publisher": {
      "@type": "Organization",
      "name": "Hercules",
      "url": "https://hercules.app"
    },
    "mainEntityOfPage": "https://hercules.app/blog/spreadsheet-error-data-risk-statistics"
  },
  {
    "@context": "https://schema.org",
    "@type": "FAQPage",
    "mainEntity": [
      {
        "@type": "Question",
        "name": "What percentage of spreadsheets contain errors?",
        "acceptedAnswer": {
          "@type": "Answer",
          "text": "A review of seven historical field audits found errors in 94% of 88 spreadsheets examined between 1995 and 2004. Historical audits by KPMG in 1998 and Coopers & Lybrand in 1997 reported errors in approximately 91% of spreadsheets they examined. These findings do not establish a current error rate for all spreadsheets."
        }
      },
      {
        "@type": "Question",
        "name": "How much do spreadsheet errors cost businesses?",
        "acceptedAnswer": {
          "@type": "Answer",
          "text": "Spreadsheet-related errors have contributed to major financial incidents. TransAlta reported a US$24 million loss in 2003. JPMorgan reported at least $6.2 billion in trading losses in 2012 amid broader risk-management failures, including spreadsheet model errors. Norway's sovereign wealth fund reported a separate benchmark-calculation error with an estimated  NOK 984 million impact."
        }
      },
      {
        "@type": "Question",
        "name": "What causes most spreadsheet errors?",
        "acceptedAnswer": {
          "@type": "Answer",
          "text": "In a 2009 audit of 50 operational spreadsheets, hard-coding, incorrect cell references, and formula logic accounted for the largest shares of recorded error instances. Hard-coding accounted for 37.7% of instances, reference errors for 32.9% of instances, and logic errors for 21.9% of instances. Some recorded instances were poor practices rather than errors that produced incorrect results."
        }
      },
      {
        "@type": "Question",
        "name": "How can businesses reduce spreadsheet errors?",
        "acceptedAnswer": {
          "@type": "Answer",
          "text": "Businesses can reduce errors by moving critical data into structured applications with validation rules, using databases instead of disconnected files, automating data entry where possible, and implementing formal review processes. The Hercules app builder lets teams build applications with configurable controls without writing code, although they must configure and test those controls correctly."
        }
      },
      {
        "@type": "Question",
        "name": "When should a business replace spreadsheets with custom software?",
        "acceptedAnswer": {
          "@type": "Answer",
          "text": "Consider replacing spreadsheets when data is shared across teams, when errors have financial or compliance consequences, when the same data is entered in multiple places, or when spreadsheets have grown too complex to audit reliably. The Hercules app builder lets businesses build custom applications for these workflows, with appropriate validation, permissions, and testing."
        }
      }
    ]
  }
]
```