The story is available at: https://samililja.wordpress.com/2014/03/28/what-if-supermarket/
While supermarkets are in general organised nicely, the insanity of bad system is not limited to IT.
Here is an hospital. What is the second most common name for an area? Waiting. Most common is toilet, because that’s what people need when they are waiting so much.
Why do we then do stupid design, optimising waiting?
Efficiency. We wish to improve things by using the parts more efficiently.
Ackoff says: The defining or critical features of a system are such that none of its parts have.
From customer perspective, the only thing that matters is the system. One failing part and the system has failed.
What happens when you design something without knowing its purpose? It creates a suboptimal design also known as waste.
Culture is powerful. It dictates how the organisation will reach its strategic targets.
But culture will not change unless the system is changed.
Deming has said that 95% of the variability in the performance is caused by the system and only 5% by the people. This is often misquoted; Ackoff talks about variability of performance, not performance.
Let’s test this assumption. The description of the exercise is available at: https://samililja.wordpress.com/2013/07/24/exercise-to-illustrate-demings-955-rule/
Here are some things that could cause delays or poor quality in Joe’s work. See also: https://samililja.wordpress.com/2013/07/24/exercise-to-illustrate-demings-955-rule/
What matters when improving the work? While we need good processes & tools and good people, the system (design and management of work) has the most effect.
These of course affect each others. Processes are designed by people and shaped by the system etc.
Methods are mostly processes and tools. Lowest level but they get the most attention. Measures get the second most attention. Very few organisations pay attention to their purpose.
Understanding Purpose allows organisation to use significant metrics that allow learning. Do not use metrics for targets or bonuses.
What is the purpose of a Formula 1 car? Go around the number of laps as fast as possible, obeying the laws of physics and competition rules.
Winner usually says “We had a good package”. Losers complain about parts that id not work.
How good is your package? How well you work towards the Purpose?
Purpose is not fuzzy or vague. Do not use adjectives, make it pragmatic and functional.
Purpose of Cobbler is to fix shoes. It is not “Offer world-class services for foot-wearables to the target segment of people who walk”.
We have two kinds of demand in our organizations: Value demand and Failure demand. Both need to be addressed, we just can not stop doing failure demand.
Could we make this process more Lean? Remove some roles and arrows and make it faster?
Making this process faster is not improvement. If customers demand changes, then the entire process is failure demand.
Doing failure demand faster allows only more failure demand.
Furthermore, if we get rid of this particular process, a similar process will surface some place else. Thinking that “Change is bad” will create failure demand in organization.
All Failure demand is a result of not being able to do Value demand. All Failure demand comes from System itself. And we can fix it by changing the system (design and management of work).
This requires we study demand, understand value and can organise to deliver value.
When organisations institutionalise a dysfunction, they create a process or a tool to get around the problem (“bug prioritisation meeting” or “JIRA ticketing system between developers and testers”). When this happens, the original problem is accepted and not fixed. Instead, the energy goes to discussion about the workaround.
Three things:
Seeing the system through System Conditions
Understanding Demand
Having clarity of Purpose and relevant measures to that
System is a difficult thing to see, especially if you want to know what to change and which lever to pull. Thing on this slide are System conditions, they help to see the system.
In order to improve the system we have to change at least one harmful system condition.
Good news: All system conditions are man-made, result of our thinking.
Bad news: Change requires changed thinking and that is difficult.
Studying Demand is useful. Try to find out the real user / customer demand, although there usually is a filtering layer in between (e.g. “Product management” or similar). Find out both failure demand and value demand.
Example of Demand analysis is available in Kanban book by David Anderson.
Principle for design and management of work.
Study and understand value demand.
Understand failure demand, especially things that go predictably wrong.
Design against these.
One consequence of demand analysis may be that project delivery turns out to be the wrong way to organise work.
We often put SW development in a container, a project, and try to manage it as a project. We isolate demand from delivery.
As soon as demand is isolated, we stop seeing it. How can we ever design against demand, if we do not see it?
Another finding may be that prioritisation hides demand. Need for prioritisation means that capability and demand are not in balance.
Instead of priority, how about capability allocation? If the red boxes turn out to be Failure Demand, it becomes more visible in this approach.
Doing 5% or 10% better is not going to keep you warm for long.
Why run after crumbs that fall of the table when there is a large cake within a reach?
Measures: Think from customer perspective (airplane seat example). Pro-tip: If your argument for doing something is “saving cost”, think harder!
What does it mean that a system is not a sum of its parts, but the product of their interactions?
It means that we should not look how well parts fit together. We must look how well the parts WORK together. And how well that fits to customer need. Because that’s all that matters.
Thank you very much.
Here is my twitter handle, let’s keep in touch.