The Utilization - Efficiency - Effectiveness Paradox
Flow Theory of Constraints
How does this all fit together?
The Utilization - Efficiency - Effectiveness Paradox
March 16, 2025
Or why some misconceptions prevent sustainable, human, and, at the same time, high-performance organizations?
I have quite a bit of experience in supporting organizations to become hyper-productive. But before we start - there are always a lot of coaches, managers, and sometimes consultants telling us how the world works!
Managers: We need high resource utilization everywhere! If everyone is working, then output and efficiency are optimal!
Coaches: Forget about output and efficiency - only effectiveness (outcome) counts! If you deliver the right stuff, the customer will value it and pay any price!
Consultants: It's all about efficiency (productivity) everywhere. If you reduce waste, you get more output and more return!
And now - what's right?
Myth Buster #1 - "Effectiveness vs. Efficiency" or "Outcome vs. Output"
Just two thoughts, and the problem is solved.
Yes, doing the right things is always better than doing the wrong stuff - no discussion about this. So focusing on finding the right stuff seems right!
But - if you want to deliver something that is so useful that everyone wants to pay a lot of money for it - so much that you can forget about resource utilization and efficiency - then you have to be innovative, and you have to run a lot of experiments fast! Right?
But what about the failed experiments? Are they waste or ineffective?
I came out of the internet business. In the beginning, we had no idea what the customer would really pay for. So we needed around 9 attempts (products, features) that failed so we could learn, and then the 10th was a blockbuster! Yes, after a while, we got better - so it went down to 5:1 - but we were never able to predict which of the 5 would finally win!
So it's impossible to know in advance what is right. If you try to find out in an analytical way, you are probably too late, and someone else was first. And if you are the second or third on the market, then your output is no longer useful.
If you want to win in a new market - you have to be efficient to have the power to run many experiments and find exactly the one blockbuster that pays all the bills!
Oh, and what about the definition of efficiency?
Efficiency = useful output / resources needed
It can be written like this: Efficiency = Effect / Resources Needed. Oh, and it's even in the definition - you have to be efficient to generate more effect with the same resources!
Problem solved! There is no paradox.
You have to be efficient to generate more useful output (outcome) with the same resources.
But there is more to solve.
Myth Buster #2 - Efficiency versus Utilization
What most managers learned (and it's not their fault):
How to improve efficiency! It is logical - if you increase resource utilization, then output goes up and, therefore, so does efficiency! Right?
So if everyone in an organization takes care to maintain high utilization, the overall efficiency must be optimal, right?
But in reality, we see something different. I helped a lot of organizations that had problems with output and, by definition, also with efficiency!
All these companies had one thing in common:
All managers were measured on the utilization of their teams.
I'll explain later - but in the graphic, you can see our reality. You can see the increase in output after some months when we got the companies to reduce the average load of the organization to around 60%!
So it is exactly the opposite - underloading boosts output by multiples.
If you underload your organization, then the overall output goes up, and the overall efficiency breaks all records!
At first glance, this may seem nonsensical. We usually assume that there is a linear relationship between capacity utilization across all functions and efficiency.
Theory of Constraints is the Key
As already mentioned, classic management assumes "that if every function is efficient, then the overall efficiency must be optimal" - but that is wrong!
The problem is that this is still based on the paradigm of local optimization.
This assumption was invalidated around 35 years ago by Eliyahu Goldratt (founder of the Theory of Constraints). It was replaced by:
"Every system has a subsystem that determines the performance of the overall system! This subsystem is called a 'constraint.' Only optimizations that serve to make better use of this constraint improve overall performance; all others are harmful or at least ineffective!"
There is always just one constraint!
At a specific time and for a specific system-goal combination.
Why can there only be one active constraint at any time?
If no constraint were active, the system would grow at any speed and be limited by the influx of material, money, and information - a constraint would form immediately.
It is also conceivable that there are systems with multiple constraints. In the case of 2 constraints, the system would be able to oscillate - both constraints would fight for supremacy.
In the case of 3 or more constraints, you can assume a chaotic system - the constraints change so quickly that performance drops and the system is destroyed.
Survivable complex adaptive systems, therefore, always form exactly one constraint - otherwise, they would not be viable.
If you accept the existence of only one constraint, the next steps are obvious:
- Find (define) this constraint
- Relieve the constraint of all non-value-adding activities
- Subordinate all management decisions to the bottleneck
Only when these three steps have been completed:
- Expand the constraint
- Start again at step 1, and don't let inertia become your constraint
That has a dramatic consequence:
When the constraint is not overloaded, all other subsystems must be underloaded.
It is not about optimizing all subsystems with much effort, but about focusing on the constraint - achieving the greatest benefit with very little effort!
And there is more - if you are in a knowledge-work, innovative, or even complex environment, then you have fluctuations everywhere. So you need some protective capacity to compensate for unforeseen volatility. This can differ from business to business, but as a rule of thumb, you need a minimum of 20% protective capacity to keep the constraint stable.
But when the constraint is utilized at only 80%, and all other subsystems must be even more underutilized to keep the constraint stable where it should be, then they are loaded to around 60%.
Here is a (real-life) example of a load-per-skill distribution from one of these companies with high factors of throughput increase. You can see that the constraint is at most 80% utilized. The next ones are less utilized!
Here are some more load distributions - not perfect yet, but on the way to underloading the constraint:
But what about these red triangles?
If you don't allow anyone to work on stuff that does not go through the constraint, then you see that skills are often underloaded far below 60%!
This is real waste! It is both ineffective because what they do will never be delivered and inefficient because there is too much capacity!
But what happens if you see this? Then you can ask some of these people to support the constraint!
Often, the constraint consists of just ~10 people. If you find 2-3 people to support it, you can quickly increase throughput by 20-30% at no cost. And suddenly, many other resources get to work because you are now less constrained!
And the red triangle is the source of all evil!
If you, as a manager, are measured by resource utilization, then you have to find work for your people. This means you must start projects (or pull stories in Agile).
But remember the constraint. The probability is high that these projects need at least a small part of the constraint - otherwise, it would not be a constraint!
So, each project you start to achieve higher local resource utilization overloads the constraint again. The constraint starts to exceed 80%, starts multitasking, makes mistakes, and is not available when you need it. The lead times go up, and the circle of doom starts.
So, in the end, your overall efficiency goes down!
If you want to increase the overall efficiency of your organization, you have to accept resource utilization of around 60% in normal teams - with just a little more at the constraint!
Paradox finally solved!
What does this look like in practice?
Fortunately, you don't always have to reinvent the wheel. The Theory of Constraints has been an integral part of successful companies for over 30 years (e.g., Microsoft, Amazon, Skype, Mazda, BOSCH, etc.). Depending on the situation, specific, tried-and-tested methods and implementation plans exist. Critical Chain Project Management for project environments and Drum-Buffer-Rope families for production environments are best known.
In knowledge work, Critical Chain is the "secret magic unicorn!"
Critical Chain Project Management (CCPM)
In addition to the prioritization of projects based on the constraint, the fever curves are the core element of CCPM. They are the key to a fully self-organized company where everyone knows what's critical and, because of the underload, is willing to support and cooperate!
Each day, the remaining duration of the open work packages is estimated, and the progress (X-axis) on the critical chain and the buffer consumption are calculated (Y-axis).
If more buffer is consumed than progress is generated, the project is "red," and the team must take action.
All projects are summarized in a portfolio overview and made available to everyone.
Everyone who has free resources is now encouraged to support the "red" projects whenever possible.
How to implement such magic!
Essentially, however, all TOC implementations follow this pattern:
- Emergency relief for the entire organization and, thus, also for the bottleneck. Little's Law immediately leads to a massive reduction in throughput times. Furthermore, multitasking in the bottleneck decreases, and throughput increases. The organization experiences a new sense of effectiveness and success.
- Planning is changed so that the bottleneck becomes visible, and all decisions are made on the basis of the bottleneck. The projects/orders are started in such a way that there is optimum capacity utilization in the bottleneck.
- The control system is changed so that there is only one signal for all those involved, enabling them to decide at any time which task has operational priority so that the bottleneck is never empty or overflowing. This signal is designed in such a way that it can be viewed by everyone on a daily basis and is operationally effective.
- Any type of flow disruption is detected - the same signal is used as for operational control (see 3). The elimination of faults has the highest priority and is supported by company management.
The first measure alone results in a significant increase in performance and, therefore, capacity. This is necessary to create space and energy for the change process.
And do you remember the bubble chart?
If you solve the paradox, then you can have much more efficiency and effectiveness - with very low utilization!
Want to learn more
We founded a free and open community about hyper-productivity. 1.5 GB of material, including videos and articles, is waiting for you. There are also free courses for getting started and more and more web sessions for getting in touch with experts! Our goal is to make you successful.
So become part of the community:
Join the free community of TOC experts
If you are more of an "I want to read about it first" person:
We wrote down our experience from over 30 years of Critical Chain implementations in a practical handbook - the "Book of Dolphins - Part 1 - CCPM."
You can decide on the price you want to pay - get it:
Book of Dolphins - Part 1 - Critical Chain
If you need support, is that also possible?
- We help you find the real bottleneck quickly! You may have an idea where your bottleneck is - but you are not sure - otherwise, you would have solved it already. There are also different types of bottlenecks, some of which are very hidden. Take advantage of our experience!
- We can help you choose the optimum control method! There are a variety of control methods that achieve the optimum effect depending on the situation. In complex systems, different methods may also be used depending on the subsystem. We know how to recognize which control system is best suited to which system and how to combine them!
- We help you get middle management behind the change! The key to rapid change is middle management. Only if they are behind the change can it be successful and sustainable. We have over 10 years of experience, even in large and complex organizations, in determining the sequence of steps through which such a change can be implemented quickly and safely. Instead of months, we think in weeks! We have found a way to transform resistance into energy! Let us surprise you!
(C) CC-BY-SA - based on the essay "Das Auslastungsparadoxon," 2020-02-28, Wolfram Müller, Björn Czybik