For whoever can stop the programme
What it cost
What happened to six organisations that adopted SAFe, in the terms a board measures. Every claim sourced, in one page.
Six organisations adopted the Scaled Agile Framework. This is what happened next, in the terms a board measures.
topANZ Banking Group
Rolled out from 2016 and expanded across the Australian division from 2017, under a mandate to deliver faster and underpin the bank’s mortgage business.
Five years on, its home loan approvals were among the slowest in Australia. The bank lost substantial share of a $2 trillion mortgage market. Its share price was down 17% across the chief executive’s seven-year tenure.
An industry analyst’s assessment was that this is not an agile organisation, and “quite sluggish in terms of some of the things that they’re doing.”
topCapital One
The programme set out to move beyond business-and-IT alignment, with the entire workforce trained and key roles certified.
In January 2023 the company eliminated all 1,100 of its Agile roles. The announcement made clear the practice had stayed siloed inside IT, without the business involvement the programme existed to achieve.
Scaled Agile published Capital One as a customer success story in 2017. The case study remained published after the roles were cut.
Peaksys
Adopted around 2020 to support entry into a new market.
Time from prioritising a feature to activating it for customers ran to approximately one year. Work in progress and multitasking rose. A commitment cycle that freezes priorities for a quarter could not accommodate a fast-moving product bet.
They concluded the framework was holding them back, and stopped.
topVolvo Cars
Two and a half years of transformation across an organisation employing around 10,000 people in software.
By September 2022 a department of eleven teams and just over a hundred people had dropped it. Their assessment: the quarterly planning event had become largely presentational, and the coordination artefact it produced added very little against the effort it consumed. The real work was organised elsewhere.
topFitbit
The success claimed in the vendor’s case study rests on company growth and product performance. That account is contradicted by the company’s public financial data.
Fitbit later abandoned the framework, and the person who had championed it left.
Fitbit remained listed on the vendor’s website as a success story after the framework had been abandoned.
United States Air Force
In December 2019 the Chief Software Officer issued a memorandum strongly discouraging rigid, prescriptive frameworks, naming SAFe.
Scaled Agile publicly offered free consulting to address the concerns. The USAF subsequently confirmed the memorandum’s conclusions stood, and that the framework would not be used in any form in its DevSecOps programme.
topWhat the evidence adds up to
Across every case that permits independent verification, the guide finds no evidence of lasting benefit against the time and money spent — and finds that the weaknesses of the operating models these organisations were trying to leave behind were amplified rather than reduced.
The organisations that recovered did so after stopping. Beijer Electronics left in June 2021 and improvement followed. Peaksys stopped. The Volvo department dropped it and worked more fluidly without it.
topWhat to do before committing further
You do not have to decide anything about the framework to act on this. You do have to be able to measure it, and most organisations running it cannot.
Establish three numbers, before any further spend is approved. Not opinions — numbers, with a date against them, taken now so that they mean something in six months.
- Time from decision to customer. How long between agreeing to do something and a customer having it. Measure end to end, not within a stage. Peaksys measured roughly a year and treated that single number as sufficient grounds to stop.
- Dependencies removed versus dependencies scheduled. Count both. The two produce very different cost curves and only one compounds in your favour. If the number scheduled is rising each quarter, the coordination cost is being industrialised rather than reduced.
- Spillover, quarter by quarter. Whether work committed in a planning cycle is completed in it, and whether the gap is closing or widening. Beijer Electronics watched this widen for three years before concluding the method was the cause.
Then ask the question that decides it. What would we expect to see if this were not working?
If the answer is “we would need more time, or more training, or the next version”, the programme cannot be failed by any evidence, and no number you collect will ever change the decision. That is worth knowing before you spend another year.
This page is a summary written by this site. It is derived from the 2024.8 edition of Information for decision-makers considering the SAFe framework, an independently curated, community-reviewed document licensed CC BY-SA 4.0. The guide is the source and carries the full evidence and every original reference. Where this page and the guide differ, the guide is correct.