There has been considerable buzz in the agile community about scaling approaches for a while now. I first noticed it at the 2014 Agile Alliance conference where the Scaled Agile Framework (SAFe) was all the rage. Every vendor booth at the conference seemed to reference SAFe or touted compatibility with SAFe.
Since 2014, even more consultants and vendors saw dollar signs related to SAFe and jumped on board. It is clearly a money maker and as such, it has become the recommended framework of choice for large consulting firms like McKinsey, BCG and Accenture. And most ALM tool providers such as VersionOne, Atlassian and even Microsoft have built in SAFe support to keep on selling tools.
In 2020, SAFe continues to be the most popular framework for scaling agile. And while you could debate the effectiveness and user satisfaction of SAFe, continued market domination seems likely.
What about alternative agile scaling approaches? If people don’t use SAFe, what are they using to scale? While precise numbers about adoption of scaling frameworks are difficult, I’ve scoured the interweb to get as accurate a picture as possible of the state of agile scaling approaches in 2020.
I’ve included the popular Annual State of Agile Report from CollabNet/VersionOne (now Digital.ai). But I wanted to expand beyond there by incorporating other survey data to avoid relying too much on one source. Thankfully I found a few other sources Not surprisingly, these surveys frequently don’t always align.
What follows is a summary of the findings organized by specific scaling approach. Later I will share some details about the various surveys that I used in my research.
As mentioned, SAFe is the most popular scaling approach. There are other options. Let’s take a look at the following agile scaling frameworks based on popularity and show trends over time for each.
Let’s start by looking at a few techniques that are not scaling approaches at all.
DAD is an interesting one. Developed by Scott Ambler and Mark Lines at IBM, it was launched in 2012 as a framework. Since then, Ambler and Lines have redefined it as a “toolkit” to help teams make context appropriate decisions. [See my related posts Get Your Disciplined Agile Certifications and PMIs Acquisition of DAD]
The DAD toolkit could be considered a superset of all available frameworks including Lean, Scrum, Kanban, and SAFe. Techniques like agile modeling and test driven development are included. I’m telling you, everything is in there including the kitchen sink.
So it is not Really fair to include Disciplined Agile in a list of scaling approaches – it is not. Still, many have done that. Here are the % of respondents that have responded to using Disciplined Agile Or Disciplined Agile Delivery (DAD) in the various surveys:
I’ve written extensively about why Spotify is not a scaling approach, even though many claim to use it that way. Please stop. Congratulations if you renamed your teams to Squads but please stop calling that an agile scaling framework. If you really want to be like Spotify, tackle your organization’s culture. Read more in A Checklist for Using the Spotify Model.
As you can see in the chart below, there are a lot of people that didn’t read my blog and still think Spotify is an agile scaling approach.
Scrum of Scrums as a technique has been around since 2001. The creators of Scrum, Jeff Sutherland and Ken Schwaber, are credited with creating the Scrum of Scrums as a technique for organizing multiple teams. As a technique, it is simply a meeting of representatives of the various Scrum teams used to coordinate work and clear dependencies.
I have found the Scrum of Scrums approach to be common and frequently misused. It is often just a glorified status meeting attended by project managers to report the status of various Scrum Teams. Which is ironic because the Scrum Framework has no role for project managers.
I suspect the reason that the reported numbers are so high for Scrum of Scrums usage is that they are actually using the technique. It may be a good technique for coordinating multiple teams, but it doesn’t give you anywhere near the scaling guidance provided by SAFe. I’ve yet to see much guidance for Scrum of Scrums and I would be surprised if most agile practitioners consider it a viable scaling approach.
It is also possible that people confuse Scrum of Scrums with Scrum at Scale, created by Jeff Sutherland. I noticed that rarely are results for both of these featured on the same report.
So I think when it is given as a choice on a survey of scaling methods, respondents think “yeah, we do that” and say yes when the reality is they do NOT use Scrum of Scrums for scaling.
Nonetheless, let’s take a look at survey responses over time from our 8 sources.
Enterprise Scrum was founded in 2001 by the late Mike Beedle, one of the 17 authors of the Agile Manifesto. It has a small but loyal following though the outlook for growth has been stymied by Beedle’s tragic death in March 2018.
There is used to be an Enterprise Scrum website and a list of certified Enterprise Scrum coaches. I have heard that some of the other Enterprise Scrum enthusiasts may be working to finish the book Beedle was working on and hoped to publish before his death.
As you can see from the chart below, about half the surveys show Enterprise Scrum at 0% while the remaining average about 4% adoption.
Large Scale Scrum (LeSS) was co-founded by Craig Larman and Bas Vodde sometime around 2006. Together they foster the LeSS community, support the Less.works website and they published 3 books related to scaling:
LeSS is based on single team Scrum, similar to Nexus and Scrum at Scale. One interesting thing about LeSS is the encouragement to actually de-scale rather than scale.
Oddly, the survey results for LeSS are skewed. While most surveys put adoption at about 5% of all approaches, the Status Quo surveys show them significantly higher as you can see below. I don’t know how to account for this though it could be related to sample size, or sample locations.
Nexus was created in 2015 by Ken Schwaber and Scrum.org. Schwaber was one of the co-creators of Scrum and one of the 17 authors of the Agile Manifesto. The Nexus guide was published in 2015, and a book was published in 2017: The Nexus Framework for Scaling Scrum, The: Continuously Delivering an Integrated Product with Multiple Scrum Teams.
Nexus is based on the Scrum Framework.
As you can see, the surveys Say adoptions vary from 0% to 15% of respondents saying they use Nexus.
Scrum @ Scale was introduced in 2017 by Jeff Sutherland and Scrum Alliance. Sutherland is the co-creator of Scrum and one of the 17 authors of the Agile Manifesto. I consider Scrum @ Scale a late-comer to the agile scaling party.
In fact, the only surveys that mention Scrum @ Scale are the Status Quo surveys. Which seems a little odd, especially since it was just introduced in 2017 and the 2017 Status Quo survey showed 8% of respondents using it.
My theory is that people confused Scrum of Scrums and Scrum @ Scale and didn’t know what they were answering. Consider the following page that was included in the KPMG Survey report from 2019:
On the left side there is a table showing the “most common scaled frameworks” based on the survey including SAFe, Scrum of Scrums, Spotify, LeSS and Other. On the right KPMG provided an overview of the “common Agile Scaling Frameworks” that include SAFe, LeSS, Disciplined Agile, Scrum @ Scale, and Nexus. Is it possible that the KPMG authors didn’t know the difference between the two? How else do you explain the lists of common frameworks?
I am not sure what to make of this category of roll your own. It has various names in the different surveys including “Individually Created”, “Custom”, and “Own Development”. Results vary quite a bit but directionally I think we can conclude that 10-15% of people scaling are using some of their own approaches to do it.
This last category of approaches got very few responses. I am not personally familiar with any of them nor have I heard of organizations using them to scale, outside of the surveys we have included.
The highest usage for any of these was for Team of Teams at 9% on the 2020 Status Quo (Scaled) Agile survey. My theory is that whoever answered the survey with Team of Teams had recently read retired US General Stanley McChrystal’s book of the same name. Alternatively, the respondent could have confused Scrum of Scrums with Team of Teams.
Behind Team of Teams would be Lean Management and Agile Portfolio Management which each had 3% of respondents on the CollabNet/VersionOne surveys.
I don’t know what to make of RAGE (Recipes for Agile Governance) which is owned by cPrime and has been around since 2013. From the cPrime site:
RAGE is Cprime’s framework for Agile Governance, or, the formalization and exercise of repeatable decision-making practices. It enables rapid decision making, based on lightweight artifacts, developed with minimum effort. Is applicable to any process at the Project, Program, and Portfolio level of any enterprise.
— cPrime website: https://www.cprime.com/rage/
Now that we’ve looked at the survey results by scaling approach, let’s take a look at the various surveys used. Here are the various published surveys used in this analysis, starting with the oldest.
Before doing this research, I’d never heard of this Status Quo (Scaled) Agile Survey. It is conducted by a professor at Koblenz University in collaboration with the GPM – German Association for project management, IPMA – International Project Management Association and Scrum.org. The study published in 2017 was the 3rd; previous studies were in 2012 and 2014.
The study is led by Ayelt Komus Phd, professor of organization and business informatics at Koblenz University of Applied Sciences. The survey was conducted online in both English and German from September 2016 to November 2016. The results were published in March 2017. Over 1,000 respondents from over 30 countries participated in the survey.
Learn more or download your own copy of the survey results here – 2016 Status Quo (Scaled) Agile Survey Results.
Scaling Frameworks Used
The entire survey results are interesting. My focus was on the popularity of scaling frameworks.
The CollabNet/VersionOne Annual State of Agile Report is one of the longest-running and most popular surveys on agility.
Agile Scaling Approaches Reported by VersionOne in 2018
I don’t know why, but the Scaling results published by VersionOne doesn’t add up to 100%.
cPrime followed their 2017 Scaling Agile survey with another in 2019.
Let’s take a look at scaling approaches which seem to have not changed much since the 2017 cPrime survey. The two biggest categories continued to dominate though both lost ground since 2017:
Where did the growth occur?
The KPMG 2019 Survey of Agility focused on a lot of things other than scaling. But it did include the question on scaling approaches used.
Details:
Scaling Approaches Used
Interesting, the KPMG survey in 2019 does not correspond well with either the cPrime or the VersionOne surveys. I suspect this is due to it being primarily European (55%) where Spotify may have more traction as a scaling approach. At 19%, SAFe is hardly dominating and just barely more than Scrum of Scrums at 17%. And Spotify was 9%. And the numbers don’t add up to 100%.
Some background on the survey:
Survey Results for Scaling Approaches
SAFe was still the dominant scaling approach reported by survey respondents.
The Scaled Agile Framework® continues to be the most popular scaling method cited by respondents (30% this year compared to 29% last year).
— Collabnet/VersionOne 13th Annual State of Agile Report
Let’s take a look Scaling Approaches reported in the 13th annual report compared to 12th annual report from CollabnetVersion report:
Interestingly, the % for the 13th Annual report added up to 100%. In the 12th Annual Report they only added up to 88%.
The team that produced the 2017 Status Quo survey provided an update in 2020.
Scaling Frameworks Used
The most recent survey from Dr. Ayelt Komus and his team at Koblenz university show different numbers than the others. And because respondents could choose from multiple approaches, the numbers won’t add up to 100%.
In April 2020, CollabNet VersionOne, XebiaLabs, and Arxan Technologies combined to form Digital.ai which is why the Annual State of Agile Report was rebranded. The actual report was produced by research firm Analysis.net Research (not clear if they produced all the reports or just this one. Digital.ai made no indication on whether they would discontinue the long-running annual survey.
Some background on the survey:
Survey Results for Scaling Approaches in the 14th Annual State of Agile Report
Not surprisingly, SAFe was still the dominant scaling approach reported by survey respondents.
The Scaled Agile Framework® continues to be the most popular scaling method cited by respondents (35% this year compared to 30% last year). As a percentage of all responses, SAFe® outdistances the next nearest response, Scrum of Scrums, by 19%.
–Digital.ai 14th Annual State of Agile Report
Let’s dig into the numbers and look at the change from the 13th to 14th Annual State of Agile Report:
Like the previous years report, the numbers seem to add up to 100% so respondents probably were not allowed to give multiple responses. Notable changes from the previous year include: