Updated: Dec 29, 2018
In the August session of the KanbanTO Meetup we discussed the ideas in the book “Commitment”, by Olav Maassen, Chris Matts & Chris Geary. The topic was well received by the people attending the meeting, but a common comment in the feedback forms was: “how is this related to Kanban?” This article attempts to close that gap.
Strictly speaking, Real Options and Kanban are unrelated concepts; that said, the idea of seeing the world as a collection of options with deferred commitment permeates Kanban.
Kanban’s aim is to establish a pull system where commitment is deferred. As part of achieving that, flow of work seen as a sequence of knowledge discovery steps, and at some point along that sequence there’s a “line” that represents the point after which a delivery promised has been made to a Customer: anything that crosses that line is now a commitment. By the same token, anything to the left of that line is still an option, something that could be potentially dropped (or deferred) without paying a significant penalty.
When a team engages on the discussion to find where this “commitment point” is, they often find that there are different levels of commitment in their workflow, representing different decision points where work can potentially be deferred for later. A typical pattern for that is separating the commitment to implement the work from the commitment to release it, which is know in Kanban as the “two-phase commit workflow.”
In recent years, the Kanban community has expanded the “Kanban Lens” to the left of the “commitment point”, and apply it to describe the often “fuzzy” process followed to create and develop options before they can be committed.
This process of “marshaling options” up to the commitment point needs to be managed, just in the same way that the process to deliver them does. For a quite a few years now, Kanban has recognized the existence of a specialized role for this, known as “Service Request Manager.”
A more recent development, lead by Patrick Steyaert and others, has been about modeling all those pre-commitment activities using what is now known as “upstream Kanban.”
The connection between “options thinking” and Kanban runs deeper, however.
In his book “Lessons in Agile Management: On the Road to Kanban”, David J. Anderson devotes a few chapters to various “anecdotes and epiphanies” that show how Real Options affected his thinking, eventually making its way into various aspects of Kanban guidance.
An example of this is the guidance on backlog prioritization: “the act of prioritizing a backlog into some form of stack ranked list implies some amount of crystal-ball gazing to second guess the future. […] Backlogs should remain an unordered list. Pull decisions should be made dynamically when a virtual kanban is available and is signalling capacity.”¹ That is, a backlog can be seen as a collection of options that can be selected later, deferring the decision to the appropriate time. Kanban then prefers “selection” over “prioritization.”
Another area where options thinking has inspired Kanban guidance is risk assessment and management. With backlogs seen as a collection of options (or untested assumptions), we can then classify those options according to multiple criteria for change and market risk, and use that information to delay risky items until the last possible moment².
¹ “Lessons in Agile Management”, David J. Anderson, pg. 358 ² “Lessons in Agile Management”, David J. Anderson, pg. 352