top of page

On Commitment In Kanban

Most people would rather be wrong than uncertain. When the prospect of being right is questionable, human beings would rather commit to a position and be mistaken, rather than hold a posture of uncertainty and avoid commitment altogether.

Our order of preference is: #1 – to be right, #2 – to be wrong, and #3 – to remain uncertain. Why #'s 1 and 2? Because we are addicted to the false sense of security that commitments give us.

Organizations that are part of a domain where uncertainty is high, such as knowledge work or professional services, pay a heavy price if they take a similar approach to commitments about what and when to work on customer requests.

Making too many commitments results in unpredictable delivery times. We have so much work in the system that we really have no idea when things are going to get done. There are too many potential outcomes (variability) stemming from unmanageable levels of work in progress (WIP). Unfortunately, even though customers are going to be disappointed, we feel good because many things have been committed to. That's a false sense of security.

We are emotionally attached to the idea of just saying yes! However, many ideas that initially seemed like good ideas are often discarded once we receive more information. In fact, 50% is a commonly observed discard rate! We should strive to keep customer requests optional for as long as we reasonable can. We should also limit the amount of time we spend in the care and feeding (grooming) of optional work – especially if 50% of it is likely to be discarded.

Another downside of committing too early is that we are more likely to abort the work once we have already invested time and money in starting it. Why? If you say yes to everything, or say yes too early, you’re likely agreeing to a lot of bad ideas! We become so emotionally attached to our commitments that it can be hard to let go. Choose wisely!

It is ideal to keep work optional and un-prioritized for as long as you can. Options have value in uncertain domains.

Developing options before converting them (committing) is an important risk mitigation strategy. It does cost money to develop options, but think of it as buying insurance to protect you from much larger negative consequences down the road (like starting work you did not really want to do and having to abort it at great expense). The more uncertain the environment, the more insurance you are likely to want.

Okay! So committing to start work too early (or taking too much of it on) and not keeping our options open can be a costly strategy in uncertain domains. We should aim to defer commitment and keep customer requests optional for as along as we reasonable we can!

How do we do that?

The Kanban Method encourages deferred commitment and provides a set of principles and practices that are receptive to developing an approach to managing risk that enables work to remain optional. In doing so, service delivery of customer requests is more predictable and reliable. It also helps to increase the likelihood that we are working on the right things at the right time.

Constructs within the Kanban Method that help to leverage deferred commitment are:

Upstream Kanban – a way to marshal options before the commitment is made. Clearly identifying commitment points.

Two-phase commitments: a commitment to start and a commitment to deliver. Mechanisms to funnel options (pay our insurance) before we choose to convert them.

Shaping Demand – a way to balance demand with capability by allocating capacity to certain types work and risk.

Enabling a responsible approach to the discarding of options so that we can avoid aborting after commitment.

A practice of limiting the amount of work in the system so that we do not make commitments that exceed our capacity and make our delivery unpredictable.

Visualizing the flow of work and the policies around our work – so that when we do commit, we can meet our promise to deliver.

Visually seeing the interruptions to the flow of work in a WIP limited system encourage us to have discussions and .

Follow me on Twitter @jamesdsteele

92 views0 comments


bottom of page