I heard someone describe effectiveness and efficiency as two ends of a spectrum. More of one means less of the other. It’s a tidy soundbite. It’s also wrong, and it quietly leads teams to optimize for the wrong thing.
Effectiveness and efficiency are not opposite ends of anything. They are separate variables that move independently. You can be highly efficient and completely ineffective, burning through work at low cost while delivering nothing the customer wanted. Or you can be both effective and efficient at the same time. The two are related, but not linked.
Then there’s a third variable that gets missed, and it matters just as much: predictability.
The Kanban Guide describes value optimization as finding the right balance of these three attributes: effectiveness, efficiency, and predictability so let’s define them.
Effectiveness is how good we are at getting the customer what they need when they need it. It’s about value.
Efficiency is how good we are at making use of the economic resources at our disposal. This could be staffing costs or equipment, or any of the things that we’re spending our money on. This is all about budgets and costs.
Predictability is how reliably we can say what we’ll deliver and when, then have that turn out to be true. It’s about the confidence others can place in our word.
Any change we make to the workflow is a trade-off across all three. Optimize hard for one and you can quietly damage the others. Squeeze efficiency by loading everyone to capacity and your predictability collapses, because a system running at 100% has no slack to absorb variation. Chase predictability with heavy buffers and padding and your efficiency suffers. Fixate on either and you can lose sight of effectiveness entirely, running a cheap, predictable process that ships the wrong thing.
So do we need all three? Yes. Do we need them equally at every moment? No. A startup searching for product-market fit should lean hard into effectiveness and worry about efficiency later. A mature operation running a known process might rightly prize predictability. The balance shifts with context, and the skill is knowing which one to favour right now without pretending the other two stopped mattering.
An interesting side-note is that many companies will say that they know how they want that balance to look, while optimizing their behaviour in a different way. A classic example in larger companies is the claim that they want to cut costs (efficiency), while heavily over-optimizing for predictability. The reality for them is that absolute cost is irrelevant so long as they can accurately predict what that cost will be. What they want is certainty.
What makes this more difficult is that we’re tuning a complex adaptive system, and it never holds still. The conditions that made today’s balance right will have moved by next quarter. There is no perfect setting that stays perfect; the moment you find it, the system has already shifted underneath you. This is ongoing work, not a problem you solve once.
It’s a mistake to treat any of these three as a single number to optimize in isolation. In a system that never stops moving, have have to watch patterns, not data points, and then we keep adjusting the balance.
There’s one wrinkle worth its own article. There are two different kinds of efficiency, and many places optimize for the wrong one. I dig into that in Keeping people busy.