What If #5 – Continuous Improvement Is Useless

What if our faith in continuous improvement is misplaced? What if one of the central tenets of Lean and Agile ways of working is just a delusion? Cargo culting? What if knowledge work is sufficiently unlike manufacturing that the whole idea of continuously and incrementally improving the way the work works has no payback? No point? What if, indeed, it’s useless – or even actually harmful?




For the avoidance of confusion, let me define what I mean here by “continuous improvement”. In general, I mean any change to the way we work, collectively, that makes us in any way more effective. That is, any change to a certain kind of task, or practice, which reduces the time and effort we take to get a certain kind of thing done.

For example, I saw one team swap out Planning Poker in favour of Silent Grouping, saving maybe 30 minutes * 10 team members = 5 hours for every fortnightly sprint planning session. On the face of it, this seems a small, but useful, improvement to the way the work works. But was it, really?

Some Of The Issues


If the working domain is merely complicated, as is largely true for production lines in a factory, then a standard process might make sense. Assembling e.g. a car has many steps, but those steps are largely definable. Improving any single step is fairly straightforward. Shorten the bolt to reduce the number of turns required to reach the necessary torque setting. Redesign the part(s), replacing the bolt and nut, or threaded hole, with a snap or push-fit fastener, thus speeding the operation (step) and maybe saving on parts costs too. Work in a complex domain, such as much of knowledge work, whether collaborative or not, does not much resemble this scenario.


In manufacturing, a standard process is relatively straightforward to sustain. Jobs (steps) are pretty much self-contained, and simple for new workers to pick up. Experiments (with e.g. improvements) are fairly simple to conduct. Case and effect are fairly obvious. Change a thing. See (objectively) if it made a positive different. Again, knowledge work, particularly collaborative knowledge work, is not much like this. Change a thing. And guess whether the thing you changed made any kind of difference. Or did the difference (if detectable) come from a myriad of other uncontrolled – and uncontrollable – factors?


All the time we buy into the assumption that continuous improvement is something we want, we’ll naturally spend time on trying to make it happen. Time which might perhaps be better spent on other aspects of making ourselves, or team and our organisations more effective. How many teams, organisations have any idea what continuos improvement is doing for them and their relative effectiveness? And even for those few that do, how often is continuous improvement a delusional frame, obscuring other reasons for the improvements they track?


Even when an change does bring some positive uplift to effectiveness (or even efficiency), that uplift is most often marginal. Maybe we could better use our time, effort and focus on seeking out changes that bring a significant uplift to effectiveness. These may be rarer and more difficult to find, but maybe the payback is, overall, more worth having.


Often, I’ve seen folks make a change which does bring some notional benefit, but not a benefit that translates to the better meeting of anyone’s needs. This is called out in e.g. Theory Of Constraints as “improving a non-bottleneck” and is a total waste of time, and money. The need to be seen to be improving something every day is a need in itself, of course. <wry smile>


For me, this is the biggest issue. If process is indeed one of those concepts from manual work a.k.a. manufacturing that has no place or value in knowledge work then even if we find an improvement that looks worthwhile, by what means to we “lock it in” for the future? The core premise of continuous improvement is that we build effectiveness, improvement upon small improvement, over time. By this path we become ever more effective. As a team, as a group, as an organisation, as an industry, as a society. I now question this assumption. I’ve never seen it work in practice. Not over the long haul. Humans, individually and collectively, are just too fickle, inattentive, capricious andrandom, Yes, we can continuously improve a machine. Continuously swapping out less effective parts for upgrades, like with an F1 car. But a complex adaptive social system is NOT a machine. Nothing like. And not much like a process, either.

Alternative Frames

Can we imagine an alternative to continuous improvement, as we generally understand it? How might we possibly become more effective without improving this step or that practice in the way our work works?

Just by way of an example, how about we focus instead on improving the quality of relationships in the workplace, both within teams and across teams? How about we focus on introspection and mindfulness in the hope of becoming “better” (more capable, more effective) people? How about we work on being more skilled at dialogue? How about we apply ourselves to (better) understanding some (more) of the principles underpinning the way the work works? How about we work on improving the healthy functioning of the complex adaptive systems we call “work”? How about we continuously examine our collective assumptions and their fitness to our shared goals?

These are all ways to improve, ways which lie outside the traditional scope of what we call “continuous (process) improvement”.

What if our unexamined assumptions around the value of continuous improvement are in fact a major blocker? For those of us who seek more effective knowledge-work organisations, maybe continuous improvement isn’t the most effective way to get there. I leave you with the following timeless wisdom:

“The unexamined life is not worth living.”

~ Socrates

– Bob

Further Reading

Why Continuous Improvement May Need To Be Discontinued ~ Ron Ashkenas
What’s the Problem with Continuous Improvement? ~ LeanCor article
How Continuous Improvement Went Horribly Wrong, for Some ~ Alan Nicol

Other Posts In This Occasional Series

What If #1 – No Management
What If #2 – No Process
What If #3 – No Telling
What If #4 – No Answers