Are Agilists Arrogant? On Frameworks And Goodness-Of-Fit
This post is an opinion piece. While I discuss some scientific studies, the core argument is an opinion—and opinions change. I’m sharing this post because it's useful to challenge myself and others in certain beliefs. This is why the title contains a question mark.
Have we become arrogant as Agilists? I used to believe simplicity is always better than complex solutions. So, like many others, I dislike(d) SAFe for its complexity. I used to believe frameworks were a good way to initiate and guide change, so I favored applying something known (LeSS, FAST, etc) instead of rolling your own. I don’t think I’m at all alone in this, as I see these beliefs reflected in most of the discourse around Agile. But the more I learn about teams, people, and their organizations, the more I’m inclined to see these as fairly naive beliefs that riskily verge on arrogance.
An Example of Agile Arrogance
As I’ve written before, I’m skeptical of frameworks as a way to introduce, teach, and guide change. One big issue I have with them is that they implicitly, and sometimes explicitly, say: “Stop what you’re doing right now and implement this instead. Things will get better if you do”. If I look at the discourse around Agile, this is what most of us seem to believe when we walk into organizations and tell them what roles, events, artifacts, and procedures to implement because the framework tells us to. It's what we often signal in training, workshops, and coaching. Unfortunately, we often extend this with “And you’re lazy if you don’t.” This type of thinking was recently captured well in a well-meaning comment someone made about SAFe (I have the utmost respect for the author, and I wanted to pick an example that wasn’t over-the-top as some of them are):
“If you were given the choice of an expensive pill to lose weight or an annual gym membership, many people would take the pill […] Sometimes people don’t really want to lose weight. They just want to try this fancy pill.”.
The author was talking about the Scale Agile Framework (SAFe). The problem with this argument is that it only works if you assume that the pill doesn’t work. If the pill does work, even partially, a person who is low on cash and short of time would be pretty clever to go for the pill and not the expensive, time-consuming gym membership. Moreover, there is a clear judgment in this line of reasoning about people who don’t go for the annual gym membership. While this example is one of the more modest and well-meaning ones, there are plenty of examples in the online Agile discourse with much more colorful language and much stronger judgments (“SAFe isn’t Agile!”). All of this feels quite arrogant to me.
The general dislike of complex Agile frameworks, like SAFe and Disciplined Agile Delivery (DAD), is a great microcosm for this mindset. This dislike only makes sense if we assume that simple frameworks always trump complex ones or even that complex frameworks never work. A simpler framework like Large Scale Scrum (LeSS) or vanilla Scrum is preferred over the complexity of SAFe with all its layers, roles, and hierarchy. But what evidence do we have that simpler frameworks are better than more complex ones? How do we know with certainty that SAFe is nothing more than an Agile placebo?
Data might help. First of all, customers prefer SAFe over its simpler alternatives. In the most recent State of Agile, SAFe held a 53% majority market share, followed by 28% for Scrum@Scale, 6% for LeSS, and 3% for Nexus. While SAFe may be the VHS to the Betamax offered by other approaches, VHS gained market dominance because it was cheaper and more convenient. Second, a peer-reviewed study I published recently with Daniel Russo, Ph.D. analyzed data from 4.013 Agile teams. Our analyses did not yield meaningful differences between scaling frameworks on core indicators of team effectiveness nor in terms of stakeholder satisfaction with quality, responsiveness, or release frequency. Third, the academic studies that have compared scaling frameworks conclude that no framework is best in all situations (Almeide & Espinheira, 2021, Edison, Wang & Conboy, 2022). Instead, the authors recommend that organizations pick a framework that suits their needs. This reflects the notion of goodness-of-fit.
Goodness of Fit
The notion of goodness-of-fit leads to a more nuanced conversation because it considers the size of the gap between the desired state and the organization's current state. It isn’t based on the maxim that simple is always better regardless of the gap size. Instead, what will work best fits best with what is already there.
An organization that looks at the SAFe framework may recognize more of itself and its internal processes than in something much simpler, like LeSS, FAST, or vanilla Scrum. So, the gap is much smaller. For a complex organization with complex needs—like governance, legal requirements, and regulations—a more minimalist framework like LeSS or Scrum may be assumed to lead to chaos and, thus, a huge financial risk. From that perspective, isn’t it rational to go for the less risky approach that SAFe offers?
“An organization that looks at the SAFe framework may recognize much more of itself and its internal processes in it than something much simpler, like LeSS or vanilla Scrum.”
Many Agilists cynically conclude that SAFe’s popularity is entirely due to its sales and marketing. This is only true when we assume SAFe doesn’t work at all. But the scientific evidence we have for this doesn’t seem to support such a strong claim. Could those organizations that pick SAFe see or understand something that we don’t yet see or understand? Yes, it is easier to make blanket statements like “SAFe isn’t Agile”, but perhaps the truth is more nuanced and there is something valuable to learn.
A metaphor might help. Imagine a fitness coach who believes 20,000 steps a day is the only way to live healthily. While that may be fine for an already reasonably fit person, it won’t work for a 70-year-old with bad knees. A fitness regimen needs to account for the person's current situation if it wants to be effective. If an unfit 70-year-old with bad knees must walk 20,000 steps a day, they will give up because the gap is too great. All those Agilists who boldly state that “SAFe isn’t Agile” or even “Waterfall is horrible” are acting like this fitness coach. They ignore goodness-of-fit entirely.
So what determines goodness-of-fit? Honestly, I don’t have a clear picture of this, but the following factors seem important to me:
- The size of the organization. Smaller organizations may be more easily adaptable to organize work along a simpler framework, like LeSS or Scrum.
- Existing belief systems about leadership and management. Leaders in some organizations already intuitively grasp democratic leadership and how to let teams take control and resolve problems. However, in some organizations, teams and management aren’t used to this yet.
- Existing belief systems about Agility. For me, Agile is about reducing the risks of complex work by collaborating closely with stakeholders and by releasing them frequently. While this paradigm is crystal clear to me, it often isn’t in organizations.
- Existing governance and budgeting structures. How money is allocated and monitoring how it is spent varies wildly. Some organizations are used to annual budgeting cycles at the team level. Others are much more flexible in allocating money to where it is needed.
- Domain. A company operating in a fast-moving market that relies heavily on technology and software has different needs than a financial institution responsible for ensuring that money is always available to its customers.
Considering such factors, I believe there are many cases where a complex framework like SAFe or Disciplined Agile Delivery is probably a better fit than simpler alternatives. A simple framework like Scrum (with Scrum of Scrums), Nexus, or LeSS may offer a better fit for other organizations. If you’re reading this and noticing a judgment like “Well, those organizations that use SAFe are just lazy for not aiming higher”, think of yourself as saying to a 70-year-old with bad knees that they’re just lazy for not walking 20,000 steps a day. Maybe a regimen of 2000 steps a day already makes this person healthier, and they may eventually be ready for more.
Down The Rabbithole
This line of thinking can also be extended into frameworks. I won’t develop this fully here, but I’ll offer a summary for your consideration. What if we also use a goodness-of-fit approach for something like the Scrum framework? Instead of doing it all at once, which may be a huge culture shock, why not begin with what already fits a little and expand from there? If you start requiring “Sprint Goals” in organizations not used to working with team-level goals, it will create resistance because it doesn’t fit. Here, too, I believe we often arrogantly create our own resistance (and then blame teams and their organizations for not going all the way).
Closing Words
I believe a certain arrogance permeates much of the online Agile discourse. I believe it stands in the way of its own mission. We are not ”uncovering better ways of [working] by doing it and helping others do it” by diminishing SAFe or particular approaches. Yes, it is fine to have an aesthetic preference for simpler frameworks (I certainly do). Still, until there is strong and unequivocal evidence that “simple frameworks always trump complex ones”, it remains nothing more than a belief.
In the meantime, I believe we can learn much from how the Scaled Agile Framework (SAFe) is much more successful at reaching organizations at their level. While it may be much sexier to think of ourselves as fitness coaches who spur our clients on to take 20,000 steps per day, perhaps we should accept that most organizations are still like 70-year-olds with bad knees who are already making progress if they’re at 2000 steps a day. Let's aim for goodness-of-fit and embrace more nuance. Our profession will be better for it.
