@andybrandt just published a post on "What’s wrong with Toyota fascination". As he states that his article is inspired by the wonderful Winter Agile Tunnig micro conference (by the way an incredibly brilliant idea and organization form) which took place last Saturday in beautiful Kraków, Poland and given that and the fact that I was the only person mentioning Toyota (once) and manufacturing (several times) in my talk I would like to comment on that post.
His basic statements are that there are
i) a bunch of lean guys out there - in another blog post and on twitter referred to as "“fad boys” of the Internet Web 2.0 community" - are real fan boys of Toyota and by liking Toyota so much they had the stupid idea of unreflectedly and simply copy 1:1 the Toyota Production System ("mimic Toyota’s assembly line") and were indeed so short sighted as to not realize at all that cars manufacturing is something completely different than software development. (Variability, creativity, etc.)
ii) Toyota is actually one of the lamest car producers out there without any sense of innovation and merely coming up w/ boring small and family cars
So, let us have a look at our feet to explain the "mimicing" part of his post. The thing with feet is: They are useful. And guess what - they have not been invented for human beings n the first place. They have been invented for insects, small animals, never walking around on two feet, no balancing involved etc. So, the original intention was, yes, transport of living beings, but far from what we are doing with them now.
It is clear to everybody that his feet, my feet everybody's (Spiderman being an exception?) feet are a lot different from insects feet (they don't stick, they have individual toes, etc). And that's because evolution (how creationists explain that, I don't know) changed feet by external pressure to what makes them useful tools for us, walking upright, successfully doing crazy things like high jump, long jump, 100m sprints in crazy speed, cycling, climbing, marathon, ultra marathon, Phelps' swimming style, you name it. In myself had some success in one or the other of the mentioned activities and I would never deny the success of the concept of feet.
So, all the smart people (and it's not me to blame for that) who looked at Toyota Production System at the time looked indeed at a manufacturing production system. And with good reason. Like the compiler theory guys at their time looked into industrialization of our "science" and trying to bring something like engineering practices into CS for the first time, these guys looked at a mature industry and especially into a production process that did nothing else than revolutionize cars manufacturing across the world (successfully).
And of course these "fad boys" (what respect towards colleagues lies in these words) did NOT just copy the TPS because, guess what, you simply can not do it. Instead what they did was look into some principles behind it AND some production theory AND the management theory of Demming (I guess another poor old "fad boy") and thought hard about what that means to software development in a transfer process. (When I was in school the principle of transferring an idea or concept from one domain to another made the difference between good and very good. And it didn't mean "copy" or "mimic").
They basically came up with some lean principles such as "build quality in", "Stop the line", "reduce work in progress", "apply the pull principle" etc. Some guy went as far as studying Queuing theory and ToC really seriously and applying the drum-buffer-rope to SE (which was the start of Kanban).
In the case of Kanban, David Anderson did all this with good reason: To establish a sustainable pace in the organizations in a transparent manner to withhold the vicious "can't you simply do more in less time" cycle. And he and his adopters and even a fad boy like me, did this with good success. Believe it or not.
Anyone having doubts or really going for the ugly picture of us guys having installed a software engineering assembly line, Toyota style, I invite to visit me at work and search for that. You will not meet the ugly face of 1920 taylorist assembly lines ...
And yes, we are not at the end of lean or Kanban, it is evolving and in fact that is what I like about it - we are getting smarter every year. Actually that's what I like about Kanban, that I can now start to look at SE or product development from totally different angles, I simply didn't see when I only worked in Scrum. (E.g. to completely refrain from projects and start a transformation towards features along a strategy rather than managing a timeline for a standing organization like that I'm working in is quite appealing to me: Does anyone here really care about the timeline of small features that amazon is pushing out all he time? I can easily plan marketing and communication along those features without an exact timeline but along the strategy and features derived.)
So, what we are looking at is a nice example of evolution through external pressure of ideas and principles behind production (insects) to SE (men). How could anybody have a different impression on that.
Oh, and even the manufacturing guys are going for variability. One large part of the success of Toyota (or Zara or ...) is indeed to work on minimizing set up times to be able to create very different pieces of product on the very same production line, because that simply is the art of lean to produce just-in-time what was just ordered. (The Zara case is btw indeed an incredible story of urning a hole business concept upside down by simply treating clothes and design just as perishable food in a lean way). And I'm telling you - these guys are really clever in parts, they have shown some adaptability and of course, they come up with different solutions in a different context by applying similar principles to their domain just as we come to different solutions in our domain.
The power of Kanban is actually that it enables us to learn from those principles by playing around with them. And you only can do that because there is a relatively fine grained model of your process.
As a final word on the potential power of Kanban, let's see what happens when I synchronize the Input and output cadence of a Kanban system at three weeks. (Yes, and I apply my WiP Limits and establish my pull principle). What I modeled in this special case is the good old Scrum, but already a very good one with WiP limits in place and so on. The missing roles I can define very well as additional rules. But in an Kanban environment what would happen to that special case of a 3-week-sprint Scum model is: that it won't last long, because given the special context it would plain simply modeled to something fitting the environment better. And the changes would be suggested by the actual developers who designed that Kanban system anyway. So, it's about empowerment of the people doing the actual work (another principle from lean manufacturing).
As to the second point, Toyota being quite uninspired and boring. It's all a matter of taste and indeed I never bought a Toyota, but I sure give them credit and respect for: the Prius, as the first mass market hybrid car, developed in at the time revolutionary 18 months; changing the luxury car market by introducing the Lexus; being the first company to introduce the SUV concept to the mass market instead of the military etc.; and again, changing and revolutionizing the way cars are manufactured today across the world.
I think the observations in "What’s wrong with Toyota fascination" are too simple. Additionally it can not be ignored that what Andy calls "mimicking" has been done with success in lots of companies. (Feet are good, if you like it or not.) But for sure, no one needs to like the ideas behind and I like the discussion about that. About the "fads" everybody needs to decide for himself.
Disclaimer: I do use Scrum in my company and we do so with success and love it!
blog comments powered by Disqus
