PSK and Value Stream Mapping

I've used Value Stream Mapping in the context of software development before. I find it conceptually useful, but not something I'd spend a lot of time doing. There's a lot of variability in software development - each iteration is a project in its own right, each individual work item has its own risks, assumptions, and dependencies. Spending time with a more detailed map with process duration and delays will very quickly become out of date. However, mapping the overall flow of work is possible, and if you can automate some of the data collection using issue tracking tools, can lead to visibility into process steps that can be improved.

I am wondering why has Value Stream Mapping havent really been featured that much in Agile in general and its almost non existent in PSK, although its a common feature of Lean. My guess is that PSK is more about optimizing flow while VSM appears to be static and more suitable where the processes and work are very uniform, predictable and visible, whereas VSM is harder to perform with knowledge work (or just meaningless). VSM give you static efficiencies for stages in the flow but ignores the dynamics of what is actually taking place and slowing down the work (the Queues that are forming). VSM might give encourage you to improve the wrong parts of the process thus increasing WIP and further destabilizing the flow. 

That seems a reasonable analysis, although VSM can be useful for identifying the valuable workflows around which enduring and sustainable teams may crystallize. You might also ascertain how complex a team's problem domain is likely to be, whether it is a product or service that might be in need of support, and the potential for automation.

Scrum puts the emphasis more on empirical process control under conditions of high uncertainty than on ferreting out waste. In my experience, lean-agile teams which do not share this emphasis often fall victim to local optimization.

Thomas, Thanks for the reply. I like the answer that "It becomes outdated" and it could have value if there was a way for it to be automated. I dont know of any tools (like Jira, VersionOne, Visio) that an automate this so that it could become useful. I think there are some tools in Lucid Chart where that can be modified to achieve some of this. Obviously if you are manufacturing something like a car on an assembly line with plenty of predictability and little variability, VSM's can be useful. But I want to think that it goes against the values outlined by PSK and Kanban in general as it focuses on individual process efficiencies which can cause unwanted queues to form and its not focused on user value holistically.  

Thanks Ian. 

"whether it is a product or service that might be in need of support, and the potential for automation"

Are you saying that VSM's would be useful in automation applications? Or are you saying that VSM can imply the need for Automation.

I think much of those in the Agile/Kanban integration movement have implied (like Claus Leopold, and i think Daniel Vacanti) that we sort of accept that we already are going to have low efficiency rates based on utilization at various points prefer a holistic approach instead of a localized one for increasing process efficiency. I want to think it is corollary of "size dont matter" paradigm. The same way the size of work in a flow is less relevant to the overall travel as the inefficiency of one part really plays a part in the whole delivery of value. 

Are you saying that VSM's would be useful in automation applications? Or are you saying that VSM can imply the need for Automation.

Perhaps it would be better to say that Value Stream Mapping can reveal opportunities for automation. It can be a useful technique when establishing a DevOps workstream, for example.

Thanks for the clarification and the rest of your points.