Operations, as Taught by Eli Goldratt · Lesson 3 of 8
The conflict behind the measurement
Why does every plant keep a measurement that does so much damage? Because underneath it sits a conflict every operations manager lives with: fight waste, or fight for flow. Build Goldratt's operations cloud, see why compromise can't win, and learn how a cloud is evaporated.
Daniel: “All right, tons per hour is poisoning us. So we drop it. Tomorrow.”
Viktor: “And then what? My people stand around and get paid for it? You’ll have half the mill idle by Thursday.”
Ruth: “You’re both right. That’s exactly why the measurement is still there.”
Underlying every problematic policy, a conflict
Whenever a damaging policy or measurement keeps existing, there must be significant reasons not to change it. We want to change it because it jeopardises one important need of the system. We go along with it because we fear that changing it will jeopardise another important need. So, Goldratt concluded:
Underlying every problematic policy or measurement there is a conflict: whether to change it, or to live with it.
To find the core problem, we must find that conflict.
The operations cloud
Goldratt built it in five steps, and he claimed it holds for any type of operation.
- Objective (A): manage well.
- Need (B): constantly fight to increase flow. Slow flow jeopardises the bottom line, and the flow jams in different places all the time.
- Requirement (D): don’t use local efficiencies as a measurement. Using them everywhere jams the flow: unneeded inventory, long lead times, poor due dates.
- Need (C): constantly fight to reduce waste. Waste jeopardises the bottom line too, and nobody volunteers to reduce it; every department asks for more.
- Requirement (D′): use local efficiencies as the prime measurement. A resource standing idle is a major waste, and local efficiency makes sure resources stand idle as little as possible.
The usual way out: compromise
The common way to deal with a conflict is to find a balance point that partly satisfies both needs. In operations the compromise is familiar to everyone:
“We use local efficiencies, but when something is really urgent, we override them by expediting.”
A compromise doesn’t solve the conflict. At best, it satisfies both needs partially. And a system that has lived with a conflict for years has usually found a compromise close to the best balance for its situation already. Moving it a little will probably make things worse, and even a better compromise brings only a small gain: not worth the suffering.
Pause and think: Daniel suggests a new rule: efficiency counts for 70% of the bonus and on-time delivery for 30%. What is that?
- A solution to the conflict, since both needs now count in the bonus
- An injection
- Another compromise: a new balance point on the same arrow
- A new measurement
It still asks for both requirements at once, so both needs stay partly unmet. A quantum improvement needs a way to satisfy both needs beyond expectations, which means removing the conflict, not splitting it.
How to evaporate a cloud
A good solution must satisfy both needs. As long as the requirement for one need conflicts with the requirement for the other, that is impossible. So a good solution must show that one of the requirements isn’t really necessary.
And there is only one way to do that. Every arrow in the cloud rests on assumptions: the reasons we believe the requirement is necessary to meet the need. If we invalidate one assumption, the arrow breaks, and the conflict no longer exists. That is what Goldratt called evaporating a cloud:
- We have to deal with the conflict because it stops us from fully satisfying both needs.
- The conflict is between the requirements.
- We believe each requirement is necessary because of certain assumptions.
- If we invalidate an assumption, we break the logical arrow.
- The conflict no longer exists.
Some arrows are defended with “over my dead body”. Those are the ones worth examining.
Which assumption to attack?
In the operations cloud, the arrow from “fight waste” to “use local efficiencies” rests on a conviction so deep that it feels like common sense:
A resource standing idle is a major waste.
If we could prove that it is beneficial to stop most resources, aggressively, from time to time, we would have proven this assumption invalid. And then nobody would be required to use local efficiencies as the prime measurement. That is exactly what the next lesson sets out to prove.
Check your understanding
Answer all five. Get four right to complete the lesson.
1. Why does a damaging measurement persist?
- Managers don’t notice the damage, because it shows up in other departments
- Changing it seems to jeopardise another important need: there is a conflict
- It is required by law
- Software requires it
Underlying every problematic policy is a conflict.
2. In the operations cloud, the two needs are…
- Low cost and high quality
- Sales and production
- Efficiency and utilisation of the plant’s most expensive machines
- Fight to increase flow, and fight to reduce waste
Both serve the objective: manage well.
3. “We use efficiencies, but expedite urgent orders” is…
- A compromise that partly satisfies both needs
- An injection
- The core problem, stated as a policy the whole plant follows
- Drum-Buffer-Rope
Compromise doesn’t solve the conflict.
4. A cloud is evaporated by…
- Finding the best compromise
- Choosing the more important need and letting the other one go
- Invalidating an assumption under one of its arrows
- Adding a third requirement
Break the arrow, and the conflict disappears.
5. Which assumption does Goldratt set out to disprove?
- Inventory is an asset
- A resource standing idle is a major waste
- Customers want short lead times more than low prices
- Variability can be eliminated
If it is false, local efficiencies are no longer required.