Financial Forecast Accuracy and Budget Cycle Statistics

Last updated: July 2026

How long an annual budget takes is well measured. How accurate the resulting forecasts turn out to be is not. This page collects what the research actually establishes on both, with the primary source, the year, and the sample size for every figure. It separates measured error from self-reported confidence, because most published figures on forecast accuracy are the second kind presented as the first.

How long the annual budget takes

32 days is the median annual budget cycle time. Top performers complete the cycle in 25 calendar days or fewer, bottom performers in 56 days or more, measured from setting objectives to a usable budget, weekends included. From APQC's Open Standards Benchmarking in planning and management accounting, n = 2,617 organizations, published November 2017. (APQC, via Marisa Brown, CFO.com)

8.7 weeks is the average budgeting cycle, unchanged over three years. The 2026 AFP FP&A Benchmarking Survey reports the average budget still takes nearly nine weeks, with no significant movement across three annual editions. Based on 332 finance and FP&A practitioners across 54 countries, fielded August-September 2025. (Association for Financial Professionals, 2026 FP&A Benchmarking Survey)

8.1 weeks vs 9.2 weeks: teams using structured scenario planning budget faster. Organizations practising structured scenario planning averaged 8.1 weeks against 9.2 weeks for those that do not, an 11% difference. Same survey and sample as above. AFP reports this as an association, not a demonstrated cause. (Association for Financial Professionals, 2026 FP&A Benchmarking Survey)

8% of organizations complete the annual budget in under a month. 53% take one to three months, 28% take three to six months (up from 25% the prior year), and 5% take longer still. FP&A Trends Survey 2025, n = 459 finance professionals worldwide (38% Europe, 38% North America, 8% Asia). See the methodology note below on this source. (FP&A Trends Survey 2025)

"Not changed significantly since 2017": the budget cycle is not getting shorter. The 2024 edition found 15% producing an annual budget in under a month and 57% in one to three months, and states the distribution has not moved significantly since the survey began in 2017. n = 383. (FP&A Trends Survey 2024)

How accurate forecasts are, when actually measured

Only one study in this corpus measures realized deviation between forecast and outcome rather than asking practitioners to assess themselves. It is nearly two decades old.

13%: the mean absolute deviation between forecast and actual results. In KPMG International and the Economist Intelligence Unit's Forecasting with Confidence, published September 2007, forecasts deviated from actual results by an average of 13%. This specific figure is based on the 354 respondents who were able to answer the question, not the full panel of 544 senior executives (35% Europe, 30% Americas, 29% Asia-Pacific; over 30% CFOs; 59% from companies above USD 1bn revenue). (KPMG International / Economist Intelligence Unit, September 2007)

1% of companies hit their forecast exactly over a three-year period. 22% landed within ±5% of actual results. Same study, September 2007, same panel of 544 senior executives. (KPMG International / Economist Intelligence Unit, September 2007)

What finance teams report about their own accuracy

The figures below are self-reported capability, not measured error. They are not comparable with the measured deviation above, and should not be read as accuracy rates.

39% of finance functions say they can forecast earnings to within ±5%. 43% say the same for revenue. From FSN's Agility in Planning, Budgeting and Forecasting, 2021, based on more than 500 senior finance professionals globally. Self-declared capability. (FSN, 2021)

42% → 39%: self-reported earnings accuracy has drifted down, not up. FSN's series shows the share claiming ±5% earnings accuracy falling from 42% in 2017 to 39% in 2020, with revenue accuracy roughly flat at 44% to 43%. Same report. (FSN, 2021)

Over 60% said they could not forecast revenue to within ±5%. A quarter could not forecast revenue to within 10%. FSN, The Future of Planning, Budgeting and Forecasting Survey 2016, n = 955 senior finance professionals across 23 sectors, recruited through the FSN Modern Finance Forum on LinkedIn. (FSN, 2016)

44% rate their own forecasts as "very high" or "good" in accuracy. This rises to 77% among organizations using fully driver-based models and falls to 27% for basic models, while 14% describe their accuracy as poor or their forecasts as essentially ignored. FP&A Trends Survey 2025, n = 459. This is a self-assessment, not a measurement of error. (FP&A Trends Survey 2025)

How far ahead teams can forecast

80% of companies say they cannot forecast beyond twelve months. More than half cannot see beyond six months. FSN notes that 20% could forecast beyond twelve months in 2017, and reports little progress since. FSN, 2021, more than 500 senior finance professionals. (FSN, 2021)

63% cannot forecast reliably beyond six months. 39% report accurate visibility only three months out, and a further 24% under six months. FP&A Trends Survey 2024, n = 383. (FP&A Trends Survey 2024)

How fast teams can reforecast

15% complete a forecast cycle in under two days. This is down from 18% in the prior edition, while 29% take more than ten business days. FP&A Trends Survey 2025, n = 459. (FP&A Trends Survey 2025)

46% reforecast monthly. 40% reforecast quarterly, and 7% on demand or daily, down from 8% the prior year. FP&A Trends Survey 2025, n = 459. (FP&A Trends Survey 2025)

58% of those taking more than a week to reforecast cannot forecast to within ±5%. FSN reports this as an association between reforecasting speed and self-declared accuracy. FSN, 2021. (FSN, 2021)

Rolling forecasts

43% of organizations use rolling forecasts. 2026 AFP FP&A Benchmarking Survey, n = 332 practitioners across 54 countries. (Association for Financial Professionals, 2026)

5% run a rolling or continuously adaptive annual budget. Distinct from running rolling forecasts alongside a fixed annual budget. FP&A Trends Survey 2025, n = 459. (FP&A Trends Survey 2025)

38% still plan on a "last year plus x%" basis. 53% use top-down planning and 46% report having adopted rolling forecasts. Categories are not mutually exclusive. FP&A Trends Survey 2025, n = 459. (FP&A Trends Survey 2025)

Why the accuracy figures are thin

The gap between this page's two accuracy sections is the finding, not an artefact of research.

14% of finance teams formally track their forecast accuracy. Reported by the AFP citing its own 2026 FP&A Benchmarking Survey, in an article by Jason Brisbane, April 2026. The article does not state the sample size for this specific question and the full report is gated, so this figure carries less support than the others on this page. (Association for Financial Professionals, April 2026)

If most organizations do not score their forecasts, there is little raw material for anyone to aggregate. That explains why the only measured deviation figure available dates from 2007, and why the widely circulated accuracy statistics are self-assessments. Any page claiming a current, measured corporate forecast error rate is either citing 2007 or citing nothing.

A note on these sources

Two of the recurring sources here, FSN and FP&A Trends, are self-selected panels recruited through professional communities and LinkedIn, and are sponsored by planning software vendors (Prophix and OneStream respectively). They publish their methodology and sample composition, which is why they are cited here, but they are not probability samples and likely over-represent organizations already engaged in FP&A transformation. Figures from these sources are best read as directional.

APQC and AFP run structured benchmarking programmes with disclosed sample sizes. APQC's own measure pages block automated access; the CFO.com articles cited here are signed, dated, and name APQC as the data source.

Sources

APQC, Open Standards Benchmarking, Planning and Management Accounting (via Marisa Brown, CFO.com, November 2017). Link
Association for Financial Professionals, 2026 FP&A Benchmarking Survey: Integrated Planning. Link
Association for Financial Professionals, "Your Forecast Does Not Have a Score. It Should" (April 2026). Link
KPMG International / Economist Intelligence Unit, Forecasting with Confidence (September 2007). Link
FSN, Agility in Planning, Budgeting and Forecasting (2021). Link
FSN, The Future of Planning, Budgeting and Forecasting Survey 2016. Link
FP&A Trends Survey 2025. Link
FP&A Trends Survey 2024. Link

Excluded figures

Widely circulated numbers that did not meet the sourcing bar, listed so the omissions are deliberate rather than invisible.

"The average USD 1bn company spends 25,000 person-days a year on budgeting." Traced to Harvard Business School Working Knowledge (August 2003), which itself attributes it to no one. No primary study identifiable. Frequently misquoted as person-hours.
Hackett Group "Digital World Class" figures (budget cycle 29% shorter, forecasts 57% faster, 57% less spent on planning, June 2025). Relative gaps only, with no absolute values, no disclosed sample size, no collection year, and a gated underlying report. Fails the context requirement.
"Top quartile teams achieve MAPE below 5%." Appears only on vendor and consultancy blogs with no primary study cited.
Gartner budget cycle and forecast accuracy figures. No primary figure freely accessible; all behind subscription.
Deloitte CFO Signals forecast accuracy. Verified as not measured: the survey publishes confidence scores and growth expectations, not accuracy against outcomes.
Academic management earnings forecast accuracy literature (mean absolute percentage error around 21%). Real and rigorous, but it measures public guidance issued by listed companies to the market, which is a different construct from internal budget accuracy. Mixing the two would misstate both.

Changelog

2026-07: Initial version. Twenty figures across budget cycle time, measured accuracy, self-reported accuracy, forecast horizon, reforecast speed, and rolling forecast adoption. Measured and self-reported accuracy are presented in separate sections deliberately. Six figure families reviewed and excluded, listed above.

Financial Forecast Accuracy and Budget Cycle Statistics (2026)