14 June 2009

CPI & Diminishing Returns?

I have a question for any "continuous process improvement" experts out there: is the CPI approach subject to the law of diminishing returns (each additional improvement yields smaller benefits than previous improvements)? If so, doesn't that sort of undermine the whole CPI mentality? If not, why not?

Here's what I'm thinking: If the law of diminishing returns applies to continuous process improvement efforts, that means CPI provides the biggest benefit during its initial phase, and provides smaller and smaller benefits over time. That might mean it's a good short-term strategy, but somewhat weak over the long-term... so maybe the "continuous" part is of dubious value?

I've only just started thinking about it in these terms. I suspect I'm on the right track, but I might be missing something here. I sincerely want to hear from people with CPI experience & insight - and my first round of google searches turned up precious little on the topic.


Mark said...

Actually, my experience with and understanding of CPI is more in line with your concept of imperfectionism. That is, while Perfection (capital P) is the ultimate ideal, it is much better to take one small step in the direction towards Perfect today vs trying to take a giant leap all the way there which is such a huge task that it never actually gets off the ground. Then tomorrow you can work on taking one more step.

I don't think that CPI necessarily assumes a big improvement at first and then smaller improvements over time. Rather, a steady stream of progress, headed in the right direction, is what is important.

PS - is this partly inspired by reading Zeno's Paradox again?

The Dan Ward said...

Interesting observation, Mark - thanks!

I guess I sort of assumed that the initial improvements would have a significant impact, as we go after the low-hanging fruit at first (plus the effect of early enthusiasm, etc). Over time, it'll require more and more time & effort & resources to identify & implement additional improvements.

If we assume that any given process has an ideal state, in which it is optimally efficient, at some point I think we must reach a point of diminishing returns.

It seems to me that the way around this is what Clatyon Christianson calls creative destruction. Rather than making incremental improvements to the process, we demolish it and start over with something better. But that's not an incremental improvement - it's a discontinuity, a change of direction.

I have a hard time swallowing the idea of a steady stream of small improvements, each of which conveys as much improvement as the previous one. I suspect it's more likely that each improvement is "better than the next."