Or “how do we create a rolling release plan?”
I’ve now seen several situations where, for whatever reason, an group implemented SAFe to align the work of multiple Teams, but then had a general feeling that there was too much overhead for the resultant outcome. For example, one group I worked with structured themselves as a Train to deal with dependencies only to find that they were able to eliminate dependencies between teams, and so found less and less to talk about in PI Planning events. Another group implemented a Train because everyone else in their organization had implemented a Train and they wanted to leverage that learning. They found that the basic ceremonies and roles didn’t provide sufficient value as a lot of the work was transactional in nature.
For these situations people asked “what changes can we make to improve our system?” While you might argue that if we are “truly Agile” we should be able to make whatever changes make sense, the fact was that in each case the groups concerned had seen benefits from the SAFe approach and so were concerned that they will also lose something of value if they went too far. The changes they were considering were things that, from their training, meant that they were no longer “doing SAFe”.
This is a great conversation to have.
Anything we can do to “descale” our work, to make things simpler, is something that should be encouraged.
We need to help these groups think through the approach. One way to think about this is to review the purposes of the various ceremonies, roles, and artifacts supported by SAFe, think about why we are doing them, and then decide “if we don’t have this ceremony (or role, or artifact), do we need this outcome and, if yes, how do we get to the outcome if we don’t do it the SAFe way”.
Let’s take a specific example. The purpose of a PI Planning event is to have multiple Teams get to alignment on the priorities for the next quarter (and keep roadmap up to date), work dependencies, and work risks. If we don’t have the PI Planning event, what do we need to put in place to get that outcome? One group, after having decided that this was something they still needed said “we will experiment with a biweekly feature refinement session attended by the POs (or their proxy) which will prioritize upcoming work, and update the view of expectations for the next couple of quarters. While this meeting will discover dependencies, it is expected that Teams are expected to work dependencies directly between them as they work the item.” This kind of thinking can be applied to all the ceremonies, roles, and artifacts.
We also need to make sure we are responding to our values. While the previous approach should surface some of these it is worthwhile to go back and check the results against your values. For example:
All this thinking must be done in the context of the wider system of delivery and these will often seen as constraints on your approach. For example, many organizations have systems in place to do roll up reporting of work and the initial set up of these systems were based on a SAFe style process. As you make changes you will need to ensure that the changes you put in place will still allow this roll up reporting.
One final word of warning; do not underestimate the importance of maintaining a cadence, a heartbeat, for you organization. I've seen some organizations say “if we don't have a regular quarterly planning session, then we can just do everything without a schedule.” This misses one of the greatest benefits a cadence has; it provides an heartbeat for the organization, something that can be relied on, something that drives the organization forward. See Why Do We Establish a Cadence of Events? for more on this. Even if you do not do planning on a cadence, you might still do demonstrations on a cadence to ensure you have an integrated solution at regular intervals, or a retrospectives on a cadence to ensure that you are working on improvements.
As said, these are all great conversations to have. We should be reviewing our ceremonies, roles, and artifacts to improve our effectiveness. We should be working to descale our operations. This is true no matter whether we have implemented SAFe or any other framework.