Agendashift is the approach used by Mike Burrows, based on his book Kanban from the Inside, in which he describes the values behind the Kanban Method. You can learn more by reading Mike’s post Agendashift in a nutshell. As part of his development of Agendashift, Mike has put together a values based delivery assessment, which he uses when working with teams. Again, I recommend reading Mike’s posts on using Agendashift as a coaching tool and debriefing an Agendashift survey if you are not familiar with Agendashift.
After listening to Mike talk about Agendashift at this year’s London Lean Kanban Day I began wondering how his approach could be used as part of a Strategy Deployment workshop. I was curious what would happen if I used the Agendashift assessment to trigger the conversations about the elements of the X-Matrix model. Specifically, how could it be used to identify change strategies, and the associated desired outcomes, in order to frame tactics as hypotheses and experiments. Mike and I had a few conversations, and it wasn’t long before I had the opportunity to give it a go. This is a description of how I went about it.
Assessment & Analysis
The initial assessment followed Mike’s post, with participants working through individual surveys before spending time analysing the aggregated results and discussing strengths, weaknesses, convergence, divergence and importance.
Having spent some time having rich conversations about current processes and practices, triggered by exploring various perspectives suggested by the survey prompts and scores, the teams had some good insights about what they considered to be their biggest problems worth solving and which required most focus. Getting agreement on what the key problems that need solving are can be thought of as agreeing the key strategies for change.
Thus this is where I broke away from Mike’s outline, in order to first consider strategies. I asked the participants to silently and individually come up with 2 to 3 change strategies each, resulting in around 20-30 items, which we then collectively grouped into similar themes to end up with 5-10 potential strategies. Dot voting (with further discussion) then reduced this down to the 3 key change strategies which everyone agreed with.
To give some examples (which I have simplified and generalised), we had strategies around focussing collaboration, communication, quality, product and value.
Having identified these key strategies, the teams could then consider what desired outcomes they hoped would be achieved by implementing them. By answering the questions “what would we like to see or hear?” and “what would we measure?”, the teams came up with possible ways, both qualitative and quantitative, which might give an indication of whether the strategies, and ultimately the tactics, were working.
Taking the 3 key strategies, I asked small groups of 3-5 people to consider the outcomes they hope to achieve with those strategies, and then consolidated the output. One reassuring observation from this part of the workshop was that some common outcomes emerged across all the strategies. This means that there were many-to-many correlations between them, suggesting a messy coherence, rather than a simplistic and reductionist relationship.
Some examples of outcomes (again simplified and generalised) were related to culture, responsiveness, quality, understanding and feedback.
Once we have strategies and outcomes, the next step is to create some hypotheses for what tactics might implements the strategies to achieve the outcomes. To do this I tweaked Mike’s hypothesis template, and used this one:
We believe that <change>
and will result in <outcomes>
With this template, the hypotheses are naturally correlated with both strategies and outcomes (where the outcomes already consist of both subjective observations and objective measures).
I asked each participant to come up with a single hypothesis, creating a range of options from which to begin defining experiments.
For example (vastly simplified and generalised!):
We believe that a technical practice
implements a quality related strategy
and will result in fewer defects
This as far as we got in the time available, but I hope its clear that once we have hypotheses like this we can start creating specific experiments with which to move into action, with the possibility that each hypotheses could be tested with multiple experiments.
While we didn’t formally go on to populate an X-Matrix, we did have most of the main elements in place – strategies, outcomes and tactics (if we consider tactics to be the actions required to test hypotheses) – along with the correlations between them. Although we didn’t discuss end results in this instance, I don’t believe it would take much to make those explicit, and come up with the correlations to the strategies and outcomes.
On a recent call with Mike he described Agendashift in terms of making the agenda for change explicit. I think that also nicely describes Strategy Deployment, and why I think there is a lot of overlap. Strategy Deployment makes the results, strategies, outcomes and tactics explicit, along with the correlations and coherence between them, and it seems that Agendashift is one way of going about this.