Goldratt assumed that every organization has to be inherently simple. This recognition is one of the four pillars of TOC, and to my mind the most important. It is in direct clash with the new Complex Theory when applied to human organizations.
Comment: I refer in this post only to organizations that have a goal and serve clients.
Is Inherent Simplicity just a philosophical concept without practical impact?
One of the most practical advises I got from Eli Goldratt was:
If the situation you are looking at seems too complex for you then:
You are looking on a too small subsystem – look at the larger system to see the simplicity
This is a very counter-intuitive advice. When you see complexity should you look for even more complexity? But, actually the situation is relieved when you analyze the larger system because what is important and mainly what is not important become clearer. Production shop-floor might look very complex to schedule. Only when you include the actual demand, distinguish between firm and forecasted demand, you realize what the real constraint is and only then the exploitation and subordination become apparent.
The term ‘simplicity’ needs to be clarified. There are two different definitions to ‘complexity’, which also clarifies what ‘simplicity’, the opposite, means.
- Many variables, with partial dependencies between them, impact the outcomes.
- It is difficult to predict the outcome of an action or a change in the value of a certain variable.
The second definition describes why complexity bothers us. The first one describes what seems to be the cause for the difficulty.
The term ‘partial dependency’ is what makes the interaction between variables complicated. When the variables are fully depended on each other then a formula can be developed to predict the combined outcome. When the variables are absolutely independent then again it is easy to calculate the total impact. It is when partial dependencies govern the output that makes it difficult to predict.
Examples for independent, fully depended variables, and partially depended:
- Several units of the same resource. The units are independent of each other.
- A production line where every problem stops the whole line. The line certainly works according to the pace of the slowest station, and every station is fully dependable of all the other stations in the line.
- A regular production floor with different work centers and enough space between them. Every work center is partially dependent on the previous ones to provide enough materials for processing.
When, on top of the complexity, every variable is exposed to significant variability then the overall complexity is overwhelming.
Can the performance of the organization be truly unpredictable?
You may call this state “chaos”, or just “on the verge of chaos”, point is that clients cannot tolerate such a performance. When I’m promised delivery at October 1st, 2pm and the delivery shows up on October 22nd, 6:30am – this is intolerable.
Is it possible to be on the verge of chaos internally, but still provide acceptable delivery to clients?
In order to achieve acceptable reliability organizations have to become simple enough. The initial impression of complexity is wrong because the partial dependencies are pushed down, so their impact on the deliveries is limited. The reduction of the partial dependencies is achieved by providing excess capacity and long lead-times. TOC simplifies it more effectively by using buffers and buffer management. What we get is good enough predictions of meeting due-dates and even being able to promise rapid-response to part of the market that is ready to pay more for quick supply.
Still, the use of the buffers means: the predictability is limited!
Even Inherent Simplicity cannot truly mean precise predictability! The whole idea is to determine the range of our ability to predict. When CCPM planning of a project predicts completion on June 2017, it actually means no later than June 2017. It could be completed earlier and we usually like it to be earlier, but the prediction of June 2017 is good enough.
Thus, the simplicity means predictions within an acceptable range!
Does simplicity means the solution can be described in one paragraph? I doubt whether one-paragraph on CCPM is enough to give the user the ability to judge the possible ramifications. Certainly we cannot describe the BOK of TOC in one paragraph.
Simplicity in radiating an idea means the idea is well understood. This is the meaning of “predictability” when we deal with marketing messages: we are able to predict what the reader, listener or spectator understands! Even here there is a certain range of interpretation that we have to live with.
What about the details of the solution itself? Is the solution necessarily easy to implement?
Easy and simple are not synonymous. The concepts could be simple, but the implementation might face obstacles, usually predictable obstacles, but overcoming the obstacles might be difficult. So, both simplicity and ease of implementation are highly desirable, but not always perfectly reachable.
We in TOC appreciate simplicity, but achieving it is a challenge. The requirements for truly good solutions are: Simplicity, Viability (possible to do in reality) and Effectiveness (achieving the objective).
An example illustrating the challenge:
Simplified-DBR is a simple effective solution for reliable delivery in manufacturing. However, for buffer-management to work properly we assume the net touch time is less than 10% of the production lead-time. This is a complication! A solution for manufacturing environments, where net-touch-time is longer than 10%, has been developed. It complicates the required information for buffer-management, but is effective.
I remember my professor for History of Physics, Prof. Sambursky, who explained to us:
“At all times, since ancient Greece, the scientists looked for the ONE formula that would explain everything. They always came with such a formula, and then a new discovered effect did not behave according to the formula. The formula was corrected to fit the behavior of that effect. Then more new effects contradicted the formula, and the formula started to be very cumbersome and it could not predict the behaviors of new effects. Then a new theory came with a new simple formula and the cycle went on again.”
TOC is basically simple. It strives to identify the Inherent Simplicity, come up with simple solutions, simple messages and easy implementations. But, we have, from time to time, to add something to deal with environments where a certain basic assumption is invalid. This is, to my mind, the most practically effective way to manage organizations.
Until a new simpler, yet effective, approach would emerge
Eli Schragenheim 
I’ve found that evaporating clouds can lead us to find practical solutions in a matter of minutes.
When a problem arises for the first time from diferent points of view it seems extremly complex, but when analized using the cloud with the people involved, the solutions seems too obvious even if it needs more than one injection. It is now when we are aware that simplicity exists. Fortunately TOC tools are there to help us improve every situation.
LikeLike
I think you are raising an important question. What is simplicity and what is complexity? Recently I have found an interesting article “A Pragmatic Complexity Framework” http://www.researchgate.net/publication/237080843. Apart from the text I refer to figure 1 – Complexity Components. I like this figure very much, because I can recognize Eli Goldratt’s complexity/simplicity statements in the component “Cognitive and / or Perceptual Complexity”. In case I can gain deeper insights about the system (via your helicopter view or a different approach) it will be possible to provide solutions based on the (gained) knowledge and clarification which finally leads to inherent simplicity.
While the component “computational complexity” is finally another word for “Complexity Theory”. In complexity theory we finally try to compute the behavior of non-linear dynamic systems. That’s not wrong, because I am using these insights as input and or solutions for NBRs or FRTs.
Personally I would not define an unsupplied order as a chaos state. For sure something is going wrong in the organization, but to define it as chaos is too heavy for me.
LikeLike
Thank you for the article. I just had a brief look into it. In my post I have defined the systems I like to deal with: organizations that have a goal and they serve clients (necessary to achieve the goal). I think this makes a huge difference to the whole relevancy of Complexity Theory to management of organizations.
If the example I have brought of a delay of an order is not one case in many thousands then I agree that it does not mean chaos. But, if this case is not an exception – I’d call it chaos. The definition is less important than the effect of clients viewing the performance as intolerable. I do assume that without clients there will be no organization.
LikeLike
Great post Eli. This should serve well as a great intro to what TOC truly is about.
LikeLike
Eli,
Find the following useful as well: http://lawsofsimplicity.com/
Mark
LikeLike
Eli…is it right to interpret the advice of seeking a larger system as taking a 30K feet or larger time frame perspective.
Taking 30K feet perspective would force us to see similar symptoms in the upstream and downstream parts of the system.
Taking larger time frame would force us to separate chronic from the one-offs.
LikeLike
Dear Eli,
I very much appreciate that you pay attention to complex theory in the field of organizational systems. Considering the publications in this field I often wonder if they make much sense. Some months ago I told you that I did some research in this respect and liked to discuss with you the insights I got. Below I present my tentative conclusions. I would be very interested to hear your response and of course I hope that it is of any use to you.
First I will answer your question concerning intrinsic simplicity. To my mind the TOC point of view is that organizations are systems When a system has an objective – for instance to serve clients – it necessarily has a bottleneck: a leverage point where management intervention can bring about the biggest possible improvement in the outcome of the system. Under this aspect the system is simple, meaning that it is clear how to manage it effectively.
In my little bit of research I, among other things, watched some presentations of Dave Snowden. To be honest I cannot make much sense of what he tries to say. Snowden argues that the preferred decision model in complex and chaotic systems is to act and to respond on the effects that follow – the act. The reason for that is, according to him, that in complex systems there is no causality. As cause and effect relations are so crucial in our TOC approach we can ask ourselves if we should we be alarmed by Snowdens opinions. To my mind there are several reasons that we really must not. Two of them are:
1. Snowdens claim that there is no causality in complex systems is at odds with standard complex theory. Champions in this field like Robert Sapolsky and George Ellis make it clear that causality reigns in all systems, in different ways at different levels. The point is that at some levels causal relations produce chaotic and unpredictable outcomes, whereas at other levels the chaos dissolves in regular patterns. Apparently Snowden is aware of this. In a public speech he addresses the issue that his conception of complexity differs from others and he tries to justify himself by saying that that is “just a mater of definition”. Just a moment. Using the concepts “complexity” and “chaotic systems” requires the intellectual obligation to at least pay attention to what the brightest minds in this field have produced. Sapolsky and Ellis have very interesting things to say and what they say improves our understanding of organizational systems. Of course if Snowden has good reasons he may redefine complexity. However that comes at a price. The Snowden interpretation of complexity-without-causality does not represent standard mathematical theory nor does it reflect biological and physical systems.
2. Maybe Snowdens redefinition of complexity makes sense in organizational systems. Here the TOC community should be in a good position to judge. As Snowden is a spokesman of Agile I suggest that we compare the results of Agile to TOC. I recently attended a webinar of Wolfgang Müller, who is an expert in both fields. He pointed out that agile has several advantages. The main one probably is that short feedback-cycles secure that teams stay in line with the requirements of the customer. However by doing so, agile software engineering projects for instance are in danger of losing consistency and ending up in architectural spaghetti. Moreover Müller tells hat only by applying TOC knowledge agile project management can make optimal use of an organizations project capacity. Agile project management is team oriented and very good for that. But to manage agile projects efficiently high level TOC knowledge must come into play. To answer to the question I asked at the beginning op this paragraph: I don’t see any advantage in leaving out causality in the explanation of organisational systems. Yes, Agile has advantages. However these are not dependent on complexity-without-causality. To the contrary the concept of causality appears to be necessary in in creating consistent results as well as in managing agile projects properly.
To summarize: Dave Snowden deviates from standard complexity theory, but his notion of complexity-without-causality is redundant and to my mind it doesn’t explain anything.
The TOC approach actually seems to be in accordance with the insights of George Ellis, a champion of complexity theory. It could even serve as a splendid example of what he calls top down causation. Top down causation determines the physical movements at lower levels. Ellis uses football as an example: the high level rules of the football game determine what happens on the playing field. Of course the relation is exactly the same in critical chain project management: critical chain management instructions define the physical operations in the projects department. (And we know that critical chain project management is very good at that.)
To take the parallel a little further it appears that Goldratts advise to take a broader perspective on the system when you are puzzled by its complexity seems somehow to be related with high level top down causation. And Top down causation is what we in the TOC approach very often bring into practice. Of course this doesn’t eliminate bottom up causation or causalities in subsystems, that sometimes run counter to our interventions. I think that a lot more can be learned in this respect from TOC researches and complexity theorists alike.
LikeLike
Willem, thank you very much. I appreciate your points and I share much of what you say about Snowden.
My point is that while an organization can be viewed as a system – it is a special system, so not everything in organizations are applicable to systems. Not just the Goal – also serving clients, which means the organizations needs to make the clients wish to buy. This is not typical to systems in general.
When Snowden says that in complex system there is no causality, I assume he means that its very difficult, maybe impossible, to identify the specific causes of a certain happening. It also means that even when you consider the causalities you still not able to predict the outcomes in any reasonable way. For me – this is the definition of complexity.
When you relieve the complexity – the power of causality appears – it is possible to predict, not accurately but reasonably.
Agile tries to face complexity by going ahead with what is known now and stop promising what we cannot predict. TOC looks into the variability and protects it with buffers, as well as with buffer management to dictate priorities. What we get is an “reasonable prediction” , which is not precise, but so much better than refusing to predict anything – which is not a state many organization can live with.
Top down causality means: you always start with the GOAL and you ask yourself what is necessary and sufficient to achieve it. The high level look forces you to look for the higher level goal. It changes the perspective. A constraint of a local sub-organization, could be negligible when the truly global organization is checked.
LikeLike
hi Eli(writer of the post)!
with all due respect i loved your presentation about a “working definition of simplicity” with regard to organizations of which exist to create/ serve a client.
i am currently enrolled to a diploma in marketing, on my quest to simplify my ambition in managing a “group of companies” very SIMPLY in a near future. Before i applied for the cause i was convinced of the following observations/establishments about organizations;
i. leadership in an organization is a idea of absolutely independent variables(behaviors, life purpose, unique experiences.)
ii. management is also an idea of absolutely independent variables(skill, ability, and commitment). but;
iii. effectiveness in an organization’s attempt to create a satisfied customer is a phenomenon of partially dependent variables(leadership, management, operations, and a diverse clientele)
I have learned that with organizations considering marketing practices, they truly appreciate the diversity in their clientele’s needs, wants and demand states, in their attempt to attain a preset objective(s). it seems to be a pretty much simple and viable theory.
All in all, i support your scientific observation that it makes it simpler to look at the totality of things but intellectually we can not ignore the significance of a “simplified and viable learning curve of/in organizational systems”
NB.
I hope i have sparkled some interesting intellectual observations,discoveries or considerations. I am new to your material and conversations, i believe they a very significant, in striving for a simple and effective approach that CAN BE CHECKED BY SOME UNKNOWN COMPLEXITIES!
THANK YOU!
LikeLike