Operations, as Taught by Eli Goldratt · Lesson 1 of 8
Holistic, or one problem at a time?
A steel mill with four departments, four problems and four expensive fixes that have nothing to do with each other. Goldratt's opening question: is it effective to deal with each problem in isolation, and if not, where is the one cause behind them all?
Monday, 7:30 a.m. The weekly review at Northgate Steel, a mill that rolls steel plate for shipyards and construction.
Logistics: “Finished-goods stock is up again. Over two hundred thousand tonnes in the yard. We’re buying a new planning system to get it under control.”
Customer service: “Complaints are up too. We’re reorganising the service department.”
Finance: “Payback on our last investments is running at seven years. The board wants a merger study, maybe a downsizing.”
HR: “And the departments are at each other’s throats. We’ve hired an organisational psychologist.”
Daniel Okafor, the plant manager, writes all four on the whiteboard. Four problems. Four projects. Four budgets.
Why a holistic approach?
Almost every manager agrees, in principle, that an organisation should be improved as a whole. Eli Goldratt began his teaching on operations by asking why. There are three reasons, and each one matters later:
- Changing one part of a system changes the other parts, for better or for worse. You can’t launch an improvement safely without seeing the whole system.
- There are endless things that could be improved, and only limited time, money and people. So you must find the few places where effort will count. That too needs a view of the whole.
- Many local improvements don’t improve the organisation at all. A department gets its work done faster, and the work then waits longer somewhere else.
Pause and think: look at Daniel’s whiteboard. What do the four initiatives have to do with each other?
- They are all about cost
- They all need new software
- Almost nothing: each deals with one complaint on its own merit
- They were all suggested by consultants who each know only one part of the plant
A planning system, a reorganisation, a merger study and a psychologist. Each is aimed at one undesirable effect, without considering the others. That is how almost every organisation proceeds, because it isn’t obvious how to proceed holistically.
Start with the undesirable effects, from every angle
To get a holistic view, Goldratt’s first step is to list the undesirable effects (UDEs) from several aspects of the organisation, not only from operations:
| Aspect | Undesirable effect at Northgate |
|---|---|
| Logistics | Inventories are too high |
| Service | Too many customer complaints |
| Finance | Payback periods are too long |
| People | Bad relationships between departments |
Then he asked a provocative question about the steel industry as a whole. Compare a typical mill’s finished-goods inventory, its lead time for plate, its due-date performance and its payback period, today and a hundred years ago. After a century of new furnaces, computers and management programmes, how much have these numbers really improved? Not by much. Every generation dealt with each problem in isolation, and the problems stayed.
What is the major cause?
If one cause sat behind all these effects, one change could remove most of them. Two answers are common.
The first approach: variability. Processes vary, skills vary, vendors vary. The implication: launch comprehensive efforts to reduce variability everywhere, and don’t expect big results soon. Variability also gives the initiatives a shared title, but it doesn’t change the fact that they still have little to do with each other.
The second approach: our mode of managing operations. The implication: it takes courage and consensus to change the way operations is run, but it can bring a real jump in performance in a very short time. It also tells you where reducing variability is worth the effort.
Pause and think: which approach did Goldratt set out to prove?
- That variability is the main cause, so it must be reduced everywhere first
- That variability only contributes; the main cause is the way operations is managed
- That neither matters; only investment does
Most improvement initiatives are launched as if variability were the main cause. Goldratt’s session was built to prove that variability is a contributor, and that the main cause is our mode of operation. The rest of this course follows that proof.
Looking for the connections
The intuition behind “be holistic” is that a change in one place has consequences in another: there are cause-and-effect connections between the aspects. So the holistic approach means looking for the cause-and-effect connections between the UDEs, in case they all grow from one core problem. If they do, one well-aimed initiative can replace four unrelated ones.
The tool for this is the current reality tree: a diagram of cause and effect, read from the bottom up as “if … then …”, that describes an undesirable situation and points to its core problem. It also helps get people to agree on that problem, which is the first condition for changing it.
Your turn. In the box above, write the most significant UDE you see in your own organisation from four aspects: logistics, service, finance and people. Then write the initiative currently aimed at each. How much do these initiatives have to do with each other?
Check your understanding
Answer all five. Get four right to complete the lesson.
1. Which is NOT one of the reasons for a holistic approach?
- A change in one part affects other parts
- Improvement resources are limited, so we must focus
- Many local improvements don’t improve the whole
- Every department should improve by the same amount
Focus, not equal improvement everywhere.
2. To get a holistic view, Goldratt started by…
- Listing UDEs from several aspects of the system
- Measuring every machine’s efficiency to find the weakest department
- Benchmarking competitors
- Asking for a bigger budget
Logistics, service, finance, people.
3. The typical response to four UDEs is four initiatives that…
- Reinforce each other, since they share the same people and budget
- Share one budget
- Have very little to do with each other
- Remove the core problem
Each is local, not holistic.
4. Treating variability as the main cause implies…
- A real jump in performance within weeks, once variation is measured
- Reducing variability everywhere, with little result in the short term
- Changing the measurements
- Doing nothing
Goldratt’s claim: variability contributes, but the mode of operation is the main cause.
5. A current reality tree is read…
- From the top down, “in order to … we must …”
- Left to right, by department
- By the size of each problem
- From the bottom up, “if … then …”
It connects UDEs by cause and effect to a core problem.