So I’m generally a fan of SAFe. Yes, I know – lot of folks aren’t (and we’ll talk about why later). And clearly, clients are fan of SAFe because they keep wanting to do it (and we’ll talk about the whys there later as well) – but understand it’s not just because I can get a payday that I like it.
First, and foremost, it offers a great first step off the “X” and toward organizational agility for traditional organizations. Next, because the notion of “three nested kanbans” and “periodic cross-org planning” just make sense, no matter what structure you’re doing. Finally, because it’s easier to have and manage a heterogenous org (meaning one where different parts are working in different ways) using the basic patterns of SAFe than pretty much any other org I can think of.
But I’m not here to get into the bottomless “SAFe rules / SAFE drools” debate.
What I want to talk about is what I think – and what I have experienced as a practitioner, working alongside SPCT’s and thought leaders in the SAFe world – is wrong with the way SAFe recommends you implement it. The “how” not the “what”…
SAFE’s recommended path – the famous ‘snake’ – is here:
So, in the model, we do the following things, in rough order:
Make the decision to go SAFe;
Train change agents and stand up a LACE;
Train leaders;
Identify Value Streams and define Agile Release Trains;
Create an implementation plan;
Train and prepare for the initial ART launches;
Launch the ART(s);
Coach and support the ARTs and their teams;
Launch more ARTs;
Add a Portfolio/LPM layer;
Continuously Improve.
In the real world, here are some possible problems with this:
[#1 - Decide] What, exactly, have we decided to transform to SAFe? A program? A division? An entire enterprise? Does the decider have the span of authority to make that decision, and to support and enforce the change work necessary?
[#2 - Change Agents] How are we pulling people into the LACE? What kind of people are stepping in? What’s in it for them? Do they have the capacity (work bandwidth) to learn and change themselves? Have we left enough lead time for them to gain the expertise? Or are we renting expertise by hiring in consultants, and if so are we really making sure we transfer enough in time to let them go?
[#3 - Leaders] Training leaders is typically more perfunctory than it needs to be. (Elsewhere, I’ll talk about the tie between a LACE and organizational leadership)
[#4 - Value Streams] We’re identifying value streams from the bottom up, which often leads to value streams that are minor tweaks on existing organizational structures. Without high-level clarity on product and service offerings, and on roadmaps for those products and services, it’s often difficult to define the subproducts that make them up – which is why VS mapping is one of the hardest things to do successfully in SAFe.
[#5 - Implementation Plan] Only now are we creating an implementation plan – which is driven by one key outcome: successfully implementing SAFe. Typically, the cost of that implementation isn’t mapped to the benefits – other than doing SAFe – that the organization is seeking.
[#10 - Portfolio] We’re folding a new high-level planning process into the change roadmap at a point where a lot of costs have been borne, and while there better already be clear benefits showing, there’s already change fatigue in the organization. Mix this with the fact that the leaders at the strategic planning level only dipped their toes into the process, and many change efforts stop as they get here.
[#11 - Continuous Improvement] As the wave of change enthusiasm declines, the energy needed to seriously drive ongoing change ebbs, and we get change around the edges at the lower levels – until the next wave arrives.
So how do we deal with these?
Well, what if we designed the implementation around four principles:
Leaders go first;
We start where the work starts – at the layer below strategy, where we define the products and services, and we leverage the well-defined work to better define value streams.
Plan change acknowledging ‘change fatigue.’ Change is designed to be incremental – with each increment designed to leave a permanent increment of improvement to be followed by later increments on no fixed schedule;
We focus change on the real problems and opportunities facing the organization.
What we’d get would be a roadmap that looked like this:
Decide to go SAFe. Decide what problems/opportunities need to be addressed by doing this. Make sure that the part of the organization that can impact those problem/opportunities are in the change and that their leaders are on board.
Train the leaders in SAFe; then train them in LPM and Product thinking;
Train change agents & stand up a LACE
Build a portfolio of products and services; then build out a portfolio of work to represent the Business Value Increments (BVI - I’ll explain these later) that directly connect to the problems/opportunities identified in #1. Define the sub-products that will be worked on and define value streams around those products;
Stand up ARTs within those value streams, using the well-defined roadmaps and work to simplify their intake process;
Anchor continuous improvement in these ARTs;
Expand laterally within the organization if there’s an appetite.
(I’ll do a snazzy graphic later)
How does this work incrementally?
Decide to go SAFe. Decide what problems/opportunities need to be addressed by doing this. Make sure that the part of the organization that can impact those problem/opportunities are in the change and that their leaders are on board.
If we stop here: We’ve identified the barriers to organizational improvement.
We may decide that the problems/opportunities that are priorities can’t be addressed by going to SAFe. We may decide that we don’t have control of the parts of the org that are needed to address the key problems or opportunities. So we may see if there are other problems worth dealing with that we can deal within inside the decision-makers span of control. Or to delay as we try and broaden the circle of decision-makers.Train the leaders in SAFe; then train them in LPM and Product thinking;
If we stop here: We’ve improved the tools the leadership has to plan and define work.
There’ll be opportunities for incremental improvement in the planning process, and a leadership oriented to improving the organization.Train change agents & stand up a LACE;
If we stop here: We’ve seeded the organization with people capable of and desiring to drive improvement, and given them some runway in the form of a mutually sustaining team.
Ultimately what drives change in organization is the network of committed, aligned people. We’ve stood up a group of people and seeded them into a network. This may be the most important step in the whole process.Build a portfolio of products and services; then build out a portfolio of work to represent the Business Value Increments (BVI - I’ll explain later) that directly connect to the problems/opportunities identified in #1. Define the sub-products that will be worked on and define value streams around those products;
If we stop here: We’ve improved the connection between strategy and work, clarified strategic decision-making, and strengthened the connection between decisions and value.
After standing up a network of change agents, the next most important change we can make is to improve the work intake process; by building out a portfolio process we can improve even cumbersome existing projects.Stand up ARTs within those value streams, using the well-defined roadmaps and work to simplify their intake process;
If we stop here: We’ve improved organizational alignment to the work to be done, and reduced both the delays and friction inherent in inter-organization handoffs and queuing.Anchor continuous improvement in these ARTs;
If we stop here: We’ve left behind teams of teams that own their own process and are committed to and enabled to adapt their environment to improve their delivery.Expand laterally within the organization if there’s an appetite.
That seems pretty good - it feels like a win at every step. How would you do it??