BT

Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ

Topics

Choose your language

InfoQ Homepage Articles Kanban’s service orientation agenda

Kanban’s service orientation agenda

Lire ce contenu en français

The first instalment of this series introduced some of the Kanban Method’s best-known practices through its sustainability agenda and its three values of transparency, balance, and collaboration. Typically, it is easy to find correspondences between these values and the practices and artefacts of other methods, agile or otherwise. And to teams that have already found sustainability – for example in the form of sustainable pace – the sustainability agenda doesn’t represent a very radical choice.

The service orientation agenda may come as quite a contrast. Building on the practices of the sustainability agenda, the service orientation agenda adds the values of customer focus, flow, and leadership. Individually, each of these brings some challenge; collectively, they can represent to a significant sense of direction, a much more outward-looking approach to change.

Value: Customer focus

                                     Know what your delivering, to whom, and why.

It’s one thing to agree with that mantra; it’s another to build it into every stage of the process of knowledge discovery (a very helpful metaphor for knowledge work). Borrowing from Lean practice, a good way to test the process is to trace it backwards from the moment of customer satisfaction, examining it thoroughly for its ability to anticipate customer needs and expectations.

In knowledge work, any such examination must be predicated on the assumption that internal knowledge of customer needs will be inadequate at the start of the delivery process, perhaps grossly so. Evolutionary delivery approaches go a long way towards addressing this problem, but still it is worth asking what contribution each stage of the process really makes towards addressing that knowledge problem.

Some good habits of customer focus can be reinforced by using the kanban board as the thinking tool for this kind of review. Taking the right-to-left format of the typical stand-up meeting (working backwards from the work that is closest to completion), we can ask questions like these for each column of the board or for specific work items:

  • Whose needs are explored in this stage of the process, and how?
  • Whose needs are neglected here, and what risks does that pose?
  • What do we learn in this stage that we don’t (or can’t) know earlier?
  • In what ways do the activities of this stage help us anticipate what will be needed?
  • What is still to be learned, and where are these outstanding uncertainties best dealt with?

In many systems there is (or should be) an identifiable commitment point beyond which the abandonment of work will be the cause of significant regret. Beyond this point, a fast and reliable service minimises the likelihood that precious effort will go wasted. On those occasions (rare, we hope) that it does happen, some root cause analysis may reveal some misunderstanding of the customer domain or to an unmanaged gap between customer expectations and what is feasible to deliver. These are likely to be opportunities for significant learning.

Upstream of that commitment point and closer to the work’s early origins, it’s a different story: abandoning a piece of work can feel like a victory. Here, “upstream kanban systems” – entire boards or the leftmost columns of existing boards – are devoted to the problem of qualifying work items for the later process. These upstream work items (individual features or entire projects) need to be advanced not just on their own merits but also on their value, urgency and risk profiles relative to their alternatives. When work-in-progress (WIP) is deliberately limited, these prioritisation decisions must result in some work being removed from the board whenever work arrives faster than the system is capable of handling.

In some of these upstream aspects, there are Personal Kanban1 systems that are more sophisticated than the typical organization’s project portfolio management system. At larger scales, modern upstream kanban systems treat work items as options, built on the assumption that a significant proportion of work items should never reach the commitment point. They are configured to generate reliable streams of high quality items that flow into the downstream process at the required rate.

These advanced techniques don’t replace customer collaboration; rather they support it. Kanban works here just as it does downstream – making the need for action visible, helping good decisions to be made, and capturing learning.

Value: Flow

When our work is kept visible in our working environments in ways that make its progress (or lack thereof) apparent, we soon learn to value flow. When we see flow impeded, it generates a level of discomfort that provokes tactical (symptomatic) and strategic (systemic) responses.

We can be proactive about flow. Just as with customer focus, we can use the kanban board as a thinking tool to guide an examination of the process. Again working right to left – from the moment of customer satisfaction backwards – we ask questions such as these:

  • How do work items leave this stage in the process? By what criteria do we know that they’re ready? How are those criteria expressed? How is the state change communicated?
  • Typically, how much time do work items spend in this stage? How much (if any) of that time is spent in active work as opposed to just waiting?
  • What are the most significant sources of unpredictability? In the work in or the waiting? Waiting for internal availability or for external dependencies to be resolved?
  • How much of this stage’s capacity is absorbed in rework? Or in failure demand, that arrives only because previous work failed to meet customer needs adequately?
  • How do work items arrive into this stage? How do we know that they’re ready to be worked on?

With experience comes a deeper appreciation of flow and its relationship with work-in-progress (WIP). Mathematically, they’re related by Little’s law and other key results from queuing theory; pragmatically, practitioners learn to visualize and control hidden forms of WIP and to stimulate flow in less obvious places.

This important relationship is apparent when state changes are charted over time in a cumulative flow diagram (CFD). This visualizes the delivery rate through each activity (the slope of each line), the amount of work sitting in each state or group of adjacent states (vertical height) and a guide to the corresponding lead time (horizontal width).

A cumulative flow diagram

Other visualizations and metrics help practitioners understand how timings are distributed, giving insight into both the internal workings of the process and the service experienced by the customer. Happily, Kanban bypasses the inhumane and often self-defeating pursuit of “resource efficiency”; the pursuit of flow sits very comfortably with the sustainability agenda.

Organizations do not have to become very large before end-to-end flows involve multiple services. Smooth flow then depends on WIP being managed not only within service boundaries but between and across them. It could be said that one service’s “upstream kanban” is another’s “dependency kanban”; the trick of course is to bring the two into alignment. Kanban at scale is less about adding layers of hierarchy (though hierarchy has its place in the allocation and management of risk) ­– much more it’s about helping work and knowledge flow faster across the organization.

Value: Leadership

Leadership occupies a pivotal role in the Kanban Method, bringing together practice and principle. On the one hand are the values and practices we’ve seen already (those of the sustainability and service orientation agendas); from these are generated numerous opportunities for acts of leadership. On the other hand are the remaining values and principles that describe essential disciplines of leadership; these belong to the survivability agenda, the subject of the next instalment in this series.

Interestingly, leadership was referenced explicitly in the definition of the method2, 3 only since 2012, in the form of this principle:

“Encourage acts of leadership at every level in your organization – from individual contributor to senior management”

Why go to the trouble of documenting something that previously had been taken for granted? The answer is a pragmatic one (in the sense that it is theory extracted from real-world observation): organizational change rarely happens without leadership.

Note however that we don’t insist that leadership comes from the top – we just need it to be present wherever change is needed. Where are transparency, balance and collaboration needed? Customer focus and flow? They’re needed everywhere of course, and they’re unlikely to be developed by accident.

This need for leadership becomes particularly apparent as soon as change needs to span functional boundaries. When these boundaries are strong, reaching out across them requires a certain effort of will, perhaps even courage! An alternative strategy might be escalate upwards, but how often does this result in genuine collaboration? A defensive response seems more likely.

The exercise and nurturing of leadership is illustrated in this Toyota-inspired leadership routine. Imagine being asked these three questions by a senior executive at your usual place of work:

  1. What is the process?
  2. How can we see that it is working?
  3. How is it improving?

The repeated and consistent use of this very deliberate leadership style is designed to keep the organisation, its products and its processes moving forwards whilst the same time growing the next generation of leaders. Even if routines such as this seem alien, it does illustrate something of the at-every-level kind of leadership that Kanban promotes, one that wants to see people and process developing in tandem.

Implementing the service orientation agenda

Service orientation with Kanban starts with viewing the organization through the Kanban “lens” of service delivery, work flow and knowledge discovery. Typically, this begins with the identification of customer needs, expectations and frustrations, matched to the corresponding capabilities and frustrations of internal systems. The tools of the sustainability agenda (visualization, WIP controls, feedback loops, and so on) are then applied with a deliberate end-to-end emphasis, extending to the customer both upstream and downstream.

This “horizontal” approach is complemented by “vertical” approaches. These too build on the tools of the sustainability agenda. Work-in-progress may be aggregated by risk category, class of service, strategic initiative or theme, or simply visualized on a kanban board at larger granularities (“epics” or “minimum viable products” (MVPs) rather than “features” or “stories” for example). At each level, scheduling policies can be informed by economic models.

In our experience, the values and practices of customer focus, flow and leadership have the power to reinvigorate organisations that have become inward-looking, excessively process-centric, or struggling to manage effectively at scale. They are well aligned to Agile values, but nevertheless they can bring some healthy challenge to orthodoxy when needed.

The next instalment in this series will examine the third and last of the three agendas, the survivability agenda. What if all change was conducted with understanding, agreement, and respect?

References

1 Personal Kanban: Mapping Work | Navigating Life, Jim Benson and Tonianne DeMaria Barry (CreateSpace Independent Publishing Platform, 2011)

2 Kanban: Successful Evolutionary Change for Your Technology Business, David J. Anderson (Blue Hole Press, 2010)

3 The Kanban Method 

About the Author

Mike Burrows (Twitter:@asplake) is UK Director of David J. Anderson & Associates (djaa.com) and Director of Training Programs at LeanKanban Inc. His book Kanban from the Inside is due later in 2014.

Rate this Article

Adoption
Style

BT