Other Considerations

Predictability and Cynefin

Fully defining the product backlog early in a release cycle depends on most of the work being in the Cynefin framework’s Complicated domain. If the work is mostly Complex, it won’t be possible to elaborate the work fully and reliably until it’s done. Recall that a major focus of a project operating primarily in the Complex domain is conducting probes to determine the very nature of the problem that needs to be solved.

Strategies such as Barry Boehm’s spiral model (Boehm, 1988) have been suggested to investigate projects that have significant Complex aspects and convert them to Complicated before diving into full-scale work. That can be a useful approach for an organization that values predictability. However, not every Complex problem can be converted to Complicated, and work on problems that are mostly Complex will never be very predictable. If you see a project that has mostly Complex elements, ask whether predictability is even theoretically possible for that project.

Predictability and Agile culture

Predictability can be a touchy subject for Agile teams. One failure mode we’ve seen with Agile adoptions is teams that refuse to provide estimates, even after their businesses describe sensible reasons for needing them. We’ve seen more than one Agile adoption shut down for this reason.

We’ve also seen instances of Agile purists advising teams to avoid providing estimates and instead to coach their whole organization into becoming more Agile so that estimates aren’t needed. Aside from being instances of the tail trying to wag the dog, these examples amount to attempts by the development teams to dictate business strategy to the business.

One of the original values described in the Agile Manifesto is customer collaboration. If you are the customer and, rather than providing what you’re asking for, your Agile teams are insisting that your business needs to redefine itself, you might suggest a renewed focus on that particular Agile value.