Statistics · Published
Scrum and SAFe Statistics (2026)
Only 11% of Agile users rated themselves very satisfied in Digital.ai's 17th State of Agile, and SAFe's share of enterprises is disputed across sources.


Scrum and SAFe are the two frameworks most test managers and program leaders meet when several Scrum teams or SAFe Agile Release Trains have to deliver into one S/4HANA program. Public data on them is thinner than vendor marketing suggests, and the headline figures often disagree with each other. This roundup collects the numbers we could trace to a named publisher, dates each one, and flags where sources conflict.
Key takeaways
Respondents reporting closer scrutiny of the business impact and ROI of Agile
Source: Digital.ai, Oct 2025
Respondents who use SAFe at the enterprise level, down more than 50% from the prior year (2023 data)
Source: Digital.ai, Jan 2024
Agile users who call themselves "very satisfied" with their Agile practice
Source: Digital.ai, Jan 2024
Adoption: Agile is widespread, but depth varies
Agile use is common in the surveys, but the surveys measure different things, so compare them with care.
- Digital.ai's 17th edition found that 71% of respondents apply Agile somewhere in their software development lifecycle. (2023 survey, published January 2024.)
- Digital.ai's 17th report counted 42% of respondents running a hybrid model that mixes Agile with DevOps or other approaches.
- The 2025 edition put the share using hybrid, blended, or homegrown Agile models at 74%. The 42% and 74% figures use different definitions, so they should not be read as a trend.
Source: Digital.ai, Jan 2024; Digital.ai, Oct 2025
Scrum and SAFe share: the framework picture is contested
Framework questions produce the sharpest disagreements in this topic. Digital.ai's 17th edition is the last one we could confirm with framework-level figures.
Source: Digital.ai, Jan 2024
- The 17th report said just 26% of respondents used SAFe, a decrease of more than 50% from the year before.
- Organizations that built their own enterprise framework or follow none at all together accounted for 34% of respondents, eight points more than SAFe users.
- Scrum@Scale and Scrum of Scrums fell to 19%, a nine-point drop from 2022.
- The 2025 edition stopped asking team-level framework questions for the first time in the report's history, so the current framework shares rest on the 2023 data. (Scrum.org's summary of the 18th edition.)
- Historical: In the 14th report, published in 2020, SAFe was the most popular scaling method and ran 19 percent ahead of Scrum@Scale.
Satisfaction, ROI and outcomes
Practitioners are using Agile widely but are not uniformly pleased with it, and leaders are asking harder questions about payback.
Source: Digital.ai, Jan 2024
- Only 11% of those who use Agile are "very satisfied," and 33% are "somewhat satisfied." The 17th survey covered 788 respondents.
- In the 2025 edition, 76% cited increased scrutiny on the business impact and ROI of Agile.
- Some 41% said they had increased Agile spending over the previous two years.
- Nearly one-third of organizations are trying to tie Agile practices to measurable business outcomes.
Project success: historical Standish Group data
The Standish Group CHAOS series is the most cited source for project outcomes, but its methods are not open to outside review. Treat these figures as historical context.
- The 2015 CHAOS edition covered outcomes from more than 10,000 projects completed between 2011 and 2015.
- Across all sizes, the table in that analysis tallies 39% of Agile projects as successful, against 11% of waterfall projects. Waterfall projects failed at 29%, while Agile failures came to 9%.
- For large projects, the same table shows 18% of Agile projects successful versus 3% of waterfall projects.
- The 1994 CHAOS report recorded 16% project success.
- Across CHAOS editions, the pattern has held at roughly one-third successful, one-half challenged and one-fifth failed.
Source: Hastie, reproduced in an arXiv paper, 2015 CHAOS data (historical)
The 2020 CHAOS edition is less consistent. A secondary summary reports a 23% success rate and 19% failure rate for highly skilled project managers, versus 58% success and 9% failure for projects with no project manager at all. A second secondary summary gives 18% for highly skilled managers on Agile projects and 91% for projects without a manager. The two summaries disagree on both numbers, so do not quote either one without going to the original.
Source: Secondary summary of Standish Group CHAOS 2020. A conflicting secondary figure appears in thestory.is, 2020 CHAOS summary.
A 2026 academic paper notes that the Standish Group's full investigation is not made available for researchers to review, and that there is no explanation of how the sample of participants was selected.
Leadership and governance
Who drives the transformation shapes how much sponsorship a Scrum or SAFe rollout gets.
Source: Digital.ai, 17th State of Agile, Jan 2024
- Business leaders and executives (32%) lead the Agile transformation, followed by individual technical teams (31%) and CIO/CTOs (20%).
- In the 2025 edition, only 49% of organizations had guardrails in place as AI adoption outpaced oversight.
AI, testing and quality
The 2025 Digital.ai edition was built around AI. Its figures on AI use are the most-cited new numbers in the Agile space, and the governance gap is the part that matters for sign-off.
Source: Digital.ai, Oct 2025
- More than one-quarter of AI users are experimenting with agentic AI that makes decisions and coordinates work across tools.
- The 2025 edition drew on a small panel, with nearly 350 participants, primarily Agile coaches and consultants from organizations with more than 20,000 employees. That is a much smaller and more specialized base than a general developer survey, so read the AI percentages as coach-level opinion.
Developer-wide data from Stack Overflow gives a broader view of AI trust. Stack Overflow's 2025 survey found 84% of respondents use or plan to use AI tools in their development process, up from 76% in 2024. Also, 46% said they do not trust the accuracy of AI tool output, up from 31% a year earlier. Stack Overflow also reports that 51% of professional developers use AI tools daily.
Source: Stack Overflow, Jul 2025
None of the Scrum or SAFe sources above measured test coverage, defect escape rates or regression outcomes for Scrum or SAFe teams. Testing-specific claims are therefore left out of this roundup rather than inferred.
Skills, training and SAFe vendor data
Scaled Agile publishes its own survey results. These figures come from the vendor, and we label them as such.
Source: Scaled Agile, State of SAFe 2025 (vendor data)
- Overall, 68% of respondents reported a significant or some increase in employee satisfaction after their organization implemented SAFe.
- The report gathered input from 1,200 global professionals.
- The 2025 State of SAFe report's top request was to simplify the framework.
- A secondary summary cites Scaled Agile's own claim that SAFe has been adopted at 20,000 enterprises and more than two million people have received SAFe training. We could not verify these counts independently.
Where the numbers conflict
These are the disagreements to resolve before you quote a figure to a steering committee.
Source: Toptal (cPrime survey), undated; Plaky, 2026 summary; Parabol, 2026
- Standalone Scrum is reported at 56% of teams, rising to 83% once hybrid methods that include Scrum are counted.
- A summary of the 16th edition says almost nine in ten respondents use Scrum, while the same summary reports 63% preferring it as their team-level method. The 63% figure is also attributed to the 17th edition by another secondary summary.
- The SAFe share has two incompatible readings. A secondary summary of the 18th edition says 44% of respondents say their enterprise uses SAFe. Digital.ai's 17th report put it at 26%. A third source describes a rise from 37% in 2021 to an anticipated 53% by 2025, which is a projection, not a measurement.
- Sample sizes also disagree. Digital.ai's 17th press release cites 788 respondents, while a Scrum.org post says the latest report draws insights from over 3,220 business and IT professionals. The post does not say which edition it means.
- Denominators also differ. The 26% SAFe figure covers all respondents, while the 35% figure circulating in secondary summaries covers only organizations practicing scaled Agile. Always check the base before you compare two numbers.
What this means for your sign-off
- Put the year and the base on every framework number. Between 26%, 44% and a projected 53%, the SAFe figure depends on which survey and which denominator you use. A steering deck that cites one number without a source line will invite challenge.
- Plan the test cadence around your own sprint and release rhythm. The surveys do not measure SAP release testing. For the phase model that fits an ECC to S/4HANA move, see SAP Activate and the RISE with SAP Methodology, with where testing fits.
- Treat AI-assisted test work as a governance item. With only 49% of organizations reporting guardrails, add an approval gate to your sign-off. Start with the SAP AI for Practitioners guide and the AI guardrail items in the template library.
- Tie exit criteria to outcomes, not velocity. With 76% of practitioners facing ROI scrutiny, a sign-off that reports only sprint throughput will be hard to defend. Define business-outcome measures before the first release train starts.
- Fund the testing function as a standing capability. Scrum and SAFe both assume cross-team coordination, which the SAP Testing CoE operating model covers in terms of roles and funding.
- Check team health before you scale. With only 11% of Agile users very satisfied in the 17th edition, run a retrospective-based health check across teams before you add a scaling layer.
Frequently asked questions
Is Scrum the most used Agile framework?
It is usually reported that way, but the numbers depend on definitions. Standalone Scrum is reported at 56% of teams, rising to 83% when hybrid methodologies are counted. Treat both as secondary figures.
How many enterprises use SAFe?
The most reliable primary figure is that just 26% of respondents said they use SAFe, a decrease of more than 50% from last year in Digital.ai's 17th edition (2023 data). Secondary summaries of the 18th edition report 44%, so this figure is unresolved.
Do Agile projects succeed more often than waterfall projects?
In the Standish data for 2011 to 2015, the table reports 39% of Agile projects as successful against 11% of waterfall projects. This is historical, and the full Standish methodology is not open to outside review.
How many Agile practitioners use AI?
Digital.ai's 2025 edition reports that AI adoption rose from 68% to 84%. Only 49% have guardrails in place, according to the same report.
How satisfied are Agile practitioners?
In the 17th edition, 11% of Agile users said they were very satisfied and 33% said somewhat satisfied, according to Digital.ai's January 2024 release.
Has SAFe adoption been measured since 2025?
Scaled Agile's own 2025 survey reports 68% of respondents reporting an increase in employee satisfaction, but that is vendor data. We found no independent 2026 measurement of SAFe share, so check the next State of Agile edition before you rely on any current framework figure.
Sources
- Digital.ai's 18th State of Agile Report press release, Digital.ai, Oct 28, 2025
- 17th State of Agile Report press release, Digital.ai, Jan 2024
- 17th State of Agile Report (resource center), Digital.ai, Jan 2024
- 17th State of Agile (PDF), Digital.ai, Jan 2024
- 14th annual State of Agile report release, Digital.ai via DSSResources, 2020 (historical)
- Scrum.org commentary on the 18th State of Agile Report, Scrum.org, 2025
- AI-driven software delivery reshapes Agile practices worldwide, ITBrief, 2025 (secondary)
- State of SAFe Report 2025, Scaled Agile, 2025 (vendor data)
- From the SAFe Summit Stage: Solving Business Problems with SAFe, Scaled Agile, 2025 (vendor data)
- SAFe 101: A Beginner's Guide to the Scaled Agile Framework, Medium, 2026 (secondary)
- Stack Overflow 2025 Developer Survey press release, Stack Overflow, Jul 29, 2025
- 2025 Stack Overflow Developer Survey, Stack Overflow, 2025
- Project Success in Agile Development Projects, arXiv (reproduces Standish 2015 CHAOS data), 2017 (historical)
- Project Managers Fail to Help Software Projects (Standish Group Chaos 2020), Medium, 2020 (secondary)
- Chaos Report: why this study about IT project management is so unique, thestory.is, 2020 (secondary)
- The rise and fall of the Chaos report figures, ResearchGate, undated (historical)
- Predicting Agile Success: The Critical Few Factors, arXiv, 2026
- Adopting Agile: The Latest Reports About The Popular Mindset, Toptal (cPrime survey), undated
- 50+ Agile Statistics You Need to Know in 2025, Notta, 2025 (secondary)
- Agile Statistics: Verified Data on Adoption and Success Rates, Plaky, 2026 (secondary)
- 16 Amazing Agile Statistics [2026], Zippia, 2026 (secondary)
- What Are The Most Popular Agile Methodologies in 2026?, Parabol, 2026 (secondary)
- Insights from the Latest State of Agile Report 2025, CertLibrary, 2025 (secondary; projection, not measurement)
Last updated: November 18, 2026. Statistics are refreshed as new research is published.
Free SAP testing templates
Master test strategy, UAT sign-off, cutover runbook, go/no-go checklist and more.
Get the weekly newsletter
New statistics, practitioner stories and templates, once a week.
More statistics
- SAP Hypercare and Post-Go-Live Incident Statistics (2026)
- ERP ROI Statistics (2026): Sourced Numbers on Returns, Budgets and S/4HANA Migrations
- SAP Licensing and Support Cost Statistics (2026)
- SAP Security and Vulnerability Statistics (2026)
- AI Coding Assistant Statistics (2026)
- SAP Fiori and User Experience Statistics (2026)
How this page is made: statistics are gathered from published research with AI-assisted search, every figure is linked to its source, and the page is refreshed as new data is published. Found an error? Tell us.