Addressing some SAFe concerns

Quite often, I encounter concerns with the Scaled Agile Framework. When looking at the Big Picture, and what many companies make out of SAFe, these concerns are valid - and should be taken serious. On the other hand, none of these concerns couldn't be addressed by a capable SPC and a management who cares about making a meaningful difference in their ways of working. In this article, I'm looking at some of them.

SAFe- Blessing or curse?

Typical Concerns with SAFe

Before digging in: I do believe that the skills and experiences of your SAFe Practice Consultants are critical for the success of SAFe. If you rely on people who don't understand the consequences of any advice they give, you're in for a rough ride. An SPCT friend of mine recently said, "Having the wrong SPC's - or skipping the guidance of an SPC altogether - will easily cost you tenfold of what a good SPC would cost you." I agree. Good SPC's are worth their weight in gold. Unfortunately, they don't grow on trees ... but that's a different story.
Now, let's take a look at some of the concerns I typically encounter, and what we need to consider.

SAFe caters to traditional mindset

SAFe acknowledges and respects that many organizations transitioning to agile practices still have a traditional mindset. Therefore, a lot of material in SAFe is written in a way that might be more palatable to people who have a low understanding of agility. This is the idea of "meet people where they are." Yes - that's contentious, but it's by far better than affronting the very people you need in order to make a change. What happens after we met them - that depends a great deal on their motivation to change, and the ability of the SPC to lead change.

Complexity overkill

First things first - if you don't have the problems SAFe addresses, don't use it. With that out of the way, SAFe can't - and usually shouldn't - be applied completely. Much rather, when we spot a problem for which SAFe proposes a solution, we can look at what SAFe says, have a discussion on whether that's worth a try - and if it is, then go for it. That neither means that you need to do things that solve problems you don't have, nor that you have to do things the way SAFe says if that doesn't work for you. For myself, when someone comes to me and says, "We're using SAFe, and have this-or-that problem," I commonly refer to a SAFe article, and the original source pointers SAFe refers to, and take the discussion from there. To me, SAFe is a discussion starter - not the panacea for all ails.

Dependencies everywhere!

SAFe recognizes that dependencies can be a challenge in complex organizations. It provides clear guidance and practices that make dependencies visible, and then offers a way to address these dependencies systematically. In many organizations, dependencies are inevitable - but at least we know which troubles we have because of them, and can start the discussion of what to do about them. I have encountered more than once that teams found their dependencies unworkable and decided to re-organize. A great discussion to have.

Length of planning intervals

SAFe employs a cadence-based planning approach with fixed-length Planning Intervals (PIs) typically spanning 8-12 weeks. Longer planning intervals provide stability and predictability for teams at the expense of flexibility. However, you can adjust the duration of PIs based on your own context. I have seen Agile Release Trains running 3 1-week Sprints plus a 1-week IP Sprint. That even fits Scrum's old slogan of, "Working Software in 30 days or less" - at scale! Try what works for you, and have the discussion.

Top-down planning processes

Maybe we have the wrong picture in our heads, but who can blame us - when we've been conditioned to think in hierarchies? In a SAFe context, we'd like to decentralize that which makes sense - but not everything makes sense to decentralize. For example, the Product Management function in SAFe could be staffed from team Product Owners, if they have the skills and capacity to succeed in that role, so it doesn't need to be hierarchical. What we need is a clear focus on the bigger chunks of value, and align all teams around them. How this happens can vary per organization, but having each team decide by themselves - doesn't bode well. Please be aware that SAFe encourages active involvement and empowerment of teams and individuals in making decisions.

Not proper Scrum

SAFe is not a strict implementation of Scrum, but rather a framework that incorporates agile principles and practices from various sources, including Scrum. Since many teams already run Scrum and are familiar with the core concepts of Scrum, SAFe has taken advantage of that and it retains the fundamental principles of transparency, inspection, and adaptation. When I coach SAFe teams, I quite often refer to the Scrum Guide, and suggest that understanding and practicing Scrum really well helps a lot getting the scaling part right. Unfortunately, what I see quite often is that organizations have already used a highly dysfunctional Scrum for years, and now want to change that. If anything, the SAFe transformation could be an opportunity to fix the Scrum stuff that was always broken and never got addressed.

Teams become Feature Factories

That's already a dysfunction - because SAFe features should focus on value, not on maximizing quantity of work delivered. We should only have a single ART Kanban that's prioritized by value - and teams should collaborate on shared objectives, using common backlogs and cross-team events to ensure a cohesive system solution that maximizes value. While sometimes, we see that individual teams may focus on delivering their own features irrespective of value, SAFe provides mechanisms such as System Demos and Inspect and Adapt events to ensure they don't go off on a tangent and do a large quantity of low-value work.

Fallacious estimation processes

A lot could be said about SAFe's estimation process, and I have my own opinion on the guidance they give. To keep it simple: its guidance is only for people who don't know how to get started. Coaching and empiricism should lead to an approach that meets the needs of those who need the estimates. We should realize that in larger organizations, where quite often multiple stakeholders, other teams and potentially large amounts of money depend on making fairly accurate forecasts for completion, estimation may have implications that don't exist in smaller contexts. My own go-to example is that we need to absolutely know whether the feature will be ready for the Trade Fair, or whether we need to mitigate. But "Well, we might or might not be ready in time for the fair" - could lead to a major PR disaster, and is probably unacceptable.

One Continuous Improvement event every 3 months is too little, too late

While the Inspect and Adapt (I+A) event occurs at the end of each Planning Interval (PI), SAFe encourages continuous improvement throughout the PI, with both the Scrum-of-Scrums (SoS) and Retrospectives being opportunities to surface and address cross-cutting issues. I myself like to give the guidance that teams are expected to make their own changes and improvements independent of the I+A event, and even share learnings with other teams during that event. I expect teams to bring short-term issues to the SoS, and only bring issues to the I+A which require longer preparation and observation periods, and have an impact far beyond the team boundaries. Most of all, we need to consider that the I+A is a point of reflection, where we detach from the daily routine, and come together as a team-of-teams to discuss the things we need to discuss that we never got to discuss before. There's always something. Yes, we could have such events more often - if you feel that it could help, there's nothing in SAFe stopping you. Try it.

Conclusion

I agree that the Scaled Agile Framework (SAFe) has areas of interpretation and concerns. Expecting it to be a silver bullet is completely unrealistic. It does provide a structured approach for larger organizations trying to make sense of this entire "Agile" thing, though - but that only works if they have someone who understands more than just what the material says. When you can address the complexities of large-scale development, cross-team collaboration and overall alignment while continuously improving, you've succeeded with SAFe. To get there, you absolutely must customize SAFe, and use it wisely - otherwise it will become a mess, and that's why you need some folks who know what they're talking about. And I've seen that mess: The shambles take years to clean up. The people who support your change initiative need to be able to address critical concerns based on their own experience and learnings. And the people in your organization who have concerns need a time and place to voice these concerns - because only then can we truly improve.