Continuous Improvement vs. Continuous Change

October 14, 2014


I’m a little down on the notion of continuous improvement these days. It’s not that it doesn’t happen – it does…sometimes. I simply fear that it promises too much. I think part of it is the terminology. You see in the real world continuous improvement is neither truly continuous or necessarily an improvement.

To begin with, I’ve critiqued the notion of continuously improving before from the point of view that keeping any process happening full time is ludicrous. Certainly once every sprint is nowhere near continuous. I guess maybe it is continuous when you compare it to other more plan driven methods, but that’s far from continuous in my book.

No, the part that I find most objectionable is the improving part. You see, it’s misleading. It suggests to me that every change is an improvement. That every effort is a step forward, not back. And that is simply not how it works. It would be better labelled periodic experimentation, or punctuated mutation. You see, in the real world, when we change something, we never really know if it’s going to work out or not. There are 50/50 odds that the change will actually make things worse! 

Of course that’s a good thing. We learn a little either way. Hopefully.

The problem I have with continuous improvement is that it sets up an unreasonable expectation in those we sell it to. To restate the sales pitch: every change will be an improvement and they happen all the time.

If every change were really an improvement, I would be worried that I had been transported to an alternative universe. That I was being monitored by aliens. That there was a black helicopter hovering over my house. I’d be making a tin foil bunny suit. Fortunately I know what universe I’m in, that there aren’t any black helicopters over my house, and tin foil tends to chafe in the damnedest places. Not too sure about the aliens…

Fortunately, many of my efforts at improvement fail. And that’s the way it should be.

Team Genetics

September 28, 2014


Today I was listing to “The Splendid Table”, a great cooking show on NPR. They were talking about variation in growing heirloom tomatoes. Somehow, that got me thinking about agile teams (probably time to see the therapist again). It occurred to me that ideas like Agile are memes.

Richard Dawkins defined a meme as “an idea, behavior, or style that spreads from person to person within a culture.” and Agile certainly fits that definition. Agile has spread from obscurity to worldwide acceptance within 20 years. In another 20 years I fully expect that waterfall, plan driven methods will be nothing but a footnote in the history books. Despite some early prognostications to the contrary, Agile has grown at a very healthy rate over the last several years.

“Richard Dawkins invented the term ‘memes’ to stand for items that are reproduced by imitation rather than reproduced genetically.”

While much of the credit belongs to the teams that actually do the hard work of making a new process work, there is also the business that has arisen around evangelizing and introducing Agile to companies that deserves a great deal of the credit. Agile training and consulting has done a remarkable job of spreading the meme throughout the software development world.

I think there are characteristics of Agile training that have made Agile “sticky” as a meme. Much of the Scrum certification is based on plenty of hands-on exercises. Training and certification has yielded a decent business. I’m not sure if anyone has the numbers, but I’d be surprised if it wasn’t a multi-million dollar enterprise worldwide. Strangely enough, much of that spreading has been through imitation. There is no shared agenda for the training, much of it is simply imitated from trainer to trainer.

When trainers and others spread the meme they are like Johnny Appleseed sowing Agile ideas across fertile corporate soil.

Genes change with each generation, and so do ideas. They go through a mixing and blending each time they are shared. Parts of the idea are forgotten, other new ideas are grafted on. Soon the original idea is unrecognizable. I think that’s kind of what has happened with XP. Extreme Programming originally contained a collection of ideas that were very potent. Things like pair programming, continuous integration and others all served as core ideas within XP. Over time, those ideas have been co-opted and found their main expression in Scrum. Today, almost no one trains teams in XP, Scrum is the dominant process that is trained and introduced to teams.

“Memes do this through the processes of variation, mutation, competition, and inheritance, each of which influence a meme’s reproductive success.”

So too does Agile. In recent years methods like Kanban and ideas like No Estimates have arisen and are becoming a meaningful part of the software development landscape. These are evolutions of the Agile meme. Agile is evolving, I wonder where it will go next…