We’re BACK! Whew – sorry about that folks. Bill and I have had some crazy life stuff going on (just bought a house! more on that in a later post) but I’m excited to get back to writing.
To kick things off – wanted to jump into a question that’s been buzzing in my head lately. When is it time to STOP improving?
I know it’s called continuous improvement, but surely no one actually intends to fiddle and tweak their process forever right?
Aside from the obvious logical conclusion that improving one piece of a process/product/business ad-infinitum will yield decreasing returns, it also quite boring. No, it’s pretty clear that we should NOT continuously improve in a literal sense.
Instead the advice often given goes some like the 80/20 rule – capture the majority of the efficiency and then move on.
This has always struck me as the type of unhelpful advice ivory towers are known for sharing. It seems to assume we have perfect information about how much “improvement” is available and thus are able to neatly calculate when we’ve squeezed most of the juice from a given process.
In my experience that reality is much less like a pareto chart and much more like mining a vein of precious metal from a mountain. We have indicators, but are not certain of how much value is still hidden beneath the rock waiting to be released.
The most useful mental framework I’ve found to deal with the “when is enough enough” question comes from the ToC classic “The Goal”.
Theory of Constraints
The reason I find it so helpful is that it expands your view to avoid the trap of local optimization at the cost of system performance. Goldratt explains the mental model using an imaginary boy scout troop and one hapless scout named Herbie.
Consider each individual “scout” below as a component part of your larger process. The performance of the larger process depends on each it’s components. In Goldratt’s analogy, Herbie is the slowest scout and thus the limiting factor in the troop’s overall speed.
Where do we start improving?
With Herbie of course, as any increase in his speed will impact the whole troop’s performance.
How much do we improve Herbie?
This is the part I find most helpful; NOT until we hit some theoretical 80% threshold of his total potential. Instead we only improve Herbie until he is no longer the slowest scout in our troop. This means we only need to measure the success metric of the larger system to identify the current and next “Herbie”. What is the biggest limiter on throughput? Work on that first, and STOP once it’s no longer your biggest bottleneck.
What about you – are there any frameworks that you use to identify when it’s time to STOP improving?