Shared publicly  - 
 
Push vs Pull

Taiichi Ohno is the engineering genius behind the famed Toyota Production System (TPS), known under various names: Lean Manufacturing, Just-In-Time Manufacturing, and Pull Manufacturing.

I like the term "pull" because it's very simple and easy to understand in contrast to "push" manufacturing. Ohno describes it using a hill and gravity as a metaphor. In a push system, it's as if the factory sits at the top of a hill, producing cars as fast and efficiently as possible. As cars leave the factory, they roll downhill to dealers. When demand is unlimited this is not a problem, because customers buy the cars as quickly as they are made. This was the case in the days of the Ford Model-T: The demand for affordable, functional cars was virtually unlimited. But when demand shrinks, cars pile up at the dealership, and dealers have to reduce prices to the point where they, and the manufacturer, lose money.

The U.S. has a large enough population that American car-makers could survive with a push manufacturing system for many years. And in the 1950s U.S. car makers were feeling no pain.

But Japan's much-smaller post-war population had much more variable demands. A factory might need to rapidly switch from making one kind of car to another, or from making cars to trucks, or ambulances, depending on the demand.

This is why Ohno and his fellow engineers at Toyota developed the pull production model, which was driven not by the factory but by customer demand. In pull manufacturing, the dealership sits at the top of the hill, and when a customer buys or orders a car, the dealer orders a car from the factory which delivers it "just in time."

In our increasingly service-oriented economy, customer demands are increasingly variable. We are moving from a push-oriented world, where manufacturers set the agenda, to a pull-oriented world, where the agenda is set by customers.

This has profound implications for the way companies organize and manage work; implications that are only beginning to be understood in many industries.
38
30
Gregor Petri's profile photoKostas Kavidopoulos's profile photoBarry O'Gorman's profile photoGregory Esau's profile photo
27 comments
 
I think the only real issue with the "pull method" which I agree, is a much better method when it comes to supply and demand, is the lack of work that ends up being needed on the Factory side, in turn the lack of work for those workers, and less need to have them around, at times, doing nothing.

This is one of the side effects of some of the Unionization of companies. There is a "need" to create more and more vehicles (in the case of a dealer/factory exemple), just to keep the workers busy, this is a decision made by the companies, as what is the need to pay folks if they aren't producing anything. But with union requirements to have to keep them around, the push is fabricated, that the cars have to be made, just to keep them busy.
 
That's a great point, +Brad Watts .

Flexibility is going to be a very key survival mechanism over the next, oh, 20-30 years. Unions to regain their relevancy, and provide security for it's members, is going to have to become part of this puzzle.
 
All activities in the industrial societies have been organised in the "push" model which was build on commoditization of products. "Pull" models were an exception. Now communications and especially Internet changes the information handicap of the consumer and builds direct relationships with the manufacturer. Push model is still here, based on low priced commodities by exploiting cheap labour, but these are like the last days of an empire. We are in the middle of a big change and changes ain't coming smoothly...
 
This approach has had a larger impact on the economy of the world than any other concept n the last 50 years, and I would argue there is much more negative impact than positive to the world as a whole. JIT manufacturing means there is much more volatility in employment than there would be otherwise (there are additional technological factors as well). That amplifies the effect of unemployment in a recession or depression - it takes a larger uplift in spending across the board to begin the positive growth cycles associated with a recovery because companies react so "efficiently" that they increase the volatility of the recovery thereby diluting its effect. 20 years ago when JIT wasn't so widely adopted, companies maintained an inventory and therefore a workforce which had the benefit of smoothing out the recovery and accelerating it.
 
Pull and Push are also very important concepts for cloud computing. Pay as you go, metered use of computing is essentialy a move from push to pull, instead of building an inventory of computing capacity (through CAPEX or Capital Expenditure) you procure the capacity Just in Time as an operational expenditure (OPEX). Just in Time had an enormoes impact on quality too. With buffers of stock,. build just in case, you could allow for bad quality because if a certain item did not work, you simply took the next from the pile and send the faulty one back for repair. With just in time there is no stock pile, every item has to work, first time. More about manufacturing best practices and cloud computing at my old blog http://leanitmanager.blogspot.com
 
+Kevin Karlin whether you're talking about people or stuff, carrying inventory is a cost. There is never a guarantee, for any company, that a growth cycle will return or a recovery materialize. While we can probably expect future economic growth, the benefits will go to companies that manage to operate profitably during a downturn.

From "Straight from the Gut" by Jack Welch:

"one of the favorite “when all else fails” arguments heard in business. “If we take the orders out of the plan, you’ll kill morale and you’ll never be able to mobilize the business when the orders come back.” That wasn’t the first or the last time I heard desperate business teams use the argument. That reasoning falls into the same category as the other plea I often heard during tough times: “You’ve cut all the fat out. Now you’re into bone and you’ll ruin the business if we cut more.” Both arguments don’t make it. They’re both weak. Management always has a tendency to take the smallest bite of the cost apple. Inevitably, managers have to keep going back, again and again, to cut more as markets deteriorate. All this does is create more uncertainty for employees. I’ve never seen a business ruined because it reduced its costs too much, too fast. When good times come again, I’ve always seen business teams mobilize quickly and take advantage of the situation."
 
+Gregor Petri thanks for the link! You excellent point about utility computing applies to services across the board.
 
+Dave Gray Understood, and agreed at a company level. The impact at a macro-economic level is different - when you combine JIT manufacturing, with JIT analytics (ie Weekly Jobs Reports) with JIT content delivery (social media and the 24/7 news cycle) and the need for every company to analyse and react to trends before their competitors you have a highly volatile economic scenario where companies are making decisions with fewer real facts (a data point is not a trend!) and most of the reaction to these data points creates a self-fulfilling prophecy because so many people are reading tea-leaves which creates more FUD than it removes.

Thanks for your excellent points and conversation :D
 
Brad Watts and those stroking him for his comment truly miss the point of dealing with skilled workers. The difference between the German model currently in effect vs. the US which considers labor a simple quantity instead of human beings with feelings and needs.

During shortages of demand the procedure in Germany is to keep workers employed at everything from make-work to a lot of those little tasks that get skipped when running full bore. Similar examples exist in France though not as codified. The simplest reason being that if you layoff workers when there are short-to-medium-term drops in demand you are likely to lose a significant number who will pick up and move to a job where they can find immediate employment.

You have no loyalty to them, no care for their needs - there is no need for loyalty on their part, either. Management than has the task and expense of training new workers. The lower wages of a new hire - so important to beancounters - doesn't make up for the time and money wasted catching up to a competitor who didn't make the same decisions.

You may sit around and discuss this in the abstract when we're toiling to creep out from the most severe recession since the 1930's - because those alternatives don't exist for our labor force, right now - but working people don't forget. The single strongest issue in organizing a union is dignity. The relationship between labor and management which aids folks in feeling confident about who they work for follows on closely.
 
Can relate to Kevin Karlin's points. All of this 'pull' rather than 'push' makes for a more complex and uncertain existence - if what many people seek is security and predictability. And perhaps it does not make for an improved society or life experience. However it seems to me that 'pull' has arrived and is here to stay. And an individuals, corporates or national economies we need to engage differently and structure ourselves differently.
 
+Kevin Karlin very interesting point. If I understand you right you are getting into the topic of feedback loops in complex adaptive systems. It does seem that volatility is increasing. It seems to me that a big part of the problem is the traditional business mindset that emphasizes prediction rather than a deep understanding of current conditions and the ability to adapt.

We have designed our industrial-age organizations using the military as a model, but we have forgotten that armies must deal with the fog of war, where prediction is impossible and bold moves sometimes necessary. It seems that many businesses today are learning lessons that the US military learned in Vietnam -- the limitations of conventional forces and organization, no matter how well-armed or technically superior.
 
+Ed Campbell there is no question in my mind that treating workers with dignity is critically important, and will only be more so as industrial economies increasingly shift toward services. I can absolutely see the sense in using down times to catch up, retool and retrain, like halftime in a sporting event.

At the same time many manufacturing jobs are moving to other companies where labor is cheaper, and many more are being eliminated due to automation. It doesn't make sense to continue to fund jobs that aren't coming back.

It also doesn't make sense for workers to force management to keep those jobs, if the end result is to make your company less competitive and hasten its demise. If your job is becoming obsolete, the faster you retrain and move to a job that's valued by the marketplace, the better.

With regard to loyalty, I'm not so sure I agree. IBM used to make it a point of pride that they employed people for life, but to maintain market relevance they had to abandon this philosophy. Loyalty is important, but it's not the most important thing.

When skilled workers are in high demand they tend to go to the most fulfilling and highest-paying jobs. Customers will jump ship when a competing company treats customers better or offers lower prices. Yes, loyalty is important but only to a point. The health of the organization must come before the health of any one part.
 
What a terrific conversation. When life gives me a moment, I'll jump back in with some additional thoughts.
 
+Gregory Esau please do! I'm working on a blog post about this. Will post a link here when done.

And thanks everybody for weighing in, I appreciate the thoughts and conversation.
 
+Dave Gray You're spot on with the feedback loop and I think that Technology is the great enabler of them :D . This is not an indictment of technology - I'm a technologist by nature and have spent 20+ years in the database, informatics, BI arena helping companies understand what data was important and why it was important. In this new era however, as a society we have gotten used to sipping from this digital fire-hose and have forgotten there is a huge difference between data, information and knowledge. There is a tendency (especially with news agencies and the banking industry ) to take data and declare it as knowledge (and to your point inferring some projection from it) without first processing it as information, which almost always involves gathering additional relevant data points.

As a footnote, I think this is by far the best Sunday morning discussion I've had in a looong time.

I hope everyone has a wonderful Sunday, I'm heading out to wrap up some yard-work while there's still some semblance of nice weather. I'll check back in later.
 
What this conversation illustrates for me, is the gap in mental models much of the world uses for thinking about their world, and 'the' world. The industrial revolution gave us the model of mechanics as how to think of how our world worked, and this still permeates much of the organizational mindset.
As we rapidly move further into the information revolution and knowledge economy (excuse the buzzwords), the need grows ever more for mental models based on living systems and biology, where the study of evolution, rather than mechanics, should inform influence our mental models.

The different forces and trends that are shaping and reshaping not just the business world, but the world is a whole are legion.
Just even having a methodology for grasping these changes, and how it might effect any particular business, it's marketplace, it's customer base, (etc) is often too much even for the best managed businesses.
It's my belief that business (or any organization, in governments and privates sectors) need to adapt a systems thinking mindset (living systems), incorporate organizational learning, and social media technologies into the entire organizational culture.

Even this conversation reveals many layers of depth, such as employee loyalty and retaining high value skilled workers, how to understand the feedback loops, how to develop feedback loops (social media technologies can play a significant role here), how to gain and use intelligence from the field, the concepts of pull verses push (demand economy verses supply economy) and how to lead and innovate in this very rapidly changing socio-economic-technological landscape, where consumers are ever more empowered, and now ever more in debt.

That's a brief overview as to how I approach the challenges that were raised here, I'd like to get back and try address them a little more specifically.
 
Dave, that is one of the bettrr papers at capturing the entire concept, and I really like the city model to illustrate it. Thanks for the pointer.
When I get back to a laptop, I'd like to add on some thoughts on order and chaos.
 
Thanks for the prompt, +Dave Gray ! Truly appreciate your willingness for engagement and dialogue.
My own interest in "this" extends beyond professional practice to a deep rooted belief that the quality of our future depends upon our ability to adapt these organizational models and philosophies.

Which, brings me (and hopefully others) to order and chaos and the very nature of life.
This is a fertile thread for this as a theme, as at the very least yourself +Dave Gray , and presumably others here, have a much deeper understanding of living systems and better yet, are actively implementing that understanding into the DNA of functional organized entities.

As we know, order is the precursor to death in living systems. In nature, this can play out over hundreds, if not thousands of years, in human culture, and more specifically for our conversation here, the business enterprise, this can play out over a life span as short as ten years.
More interesting still, is how the 'laws' of order and chaos extend down into the micro (we as individuals), right on up through the system of systems (I loved that you used that term, Dave!) to the highest order of macro systems. Each order of systems has a life span, and there is, if we were to study this with the right sets of eyes, a correlation between length of survival and the taste for chaos.

Chaos is where life happens. (The French and American revolutions were and are great examples of high order chaos) Of critical note, is what happens when the system settles back towards a state of equilibrium. In natural systems, there is no predictive outcome, in our human culture of systems, at very large scales there are no predictive outcomes (thankfully), but as we scale down into the enterprise, there is an ability to guide the process towards a desired outcome (within a set of broader parameters).

Even a casual student of natural systems will come to know that chaos is a necessary and vital part of life, it is the very essence of life.
Yet we humans, particularly as we age, have a propensity to order.
This wasn't as harmful a mindset during the more glacial pace of cultural evolution of the past 10,000 years (I believe we were very natural systems thinkers and be-ers before the scaling of agricultural civilizations began to render order on social systems), but in the hyper evolutionary pace of today's society, this propensity to order is more likely to be fatal for both the organization and the individual (adapt or die, as more than a few business book advices wisely).

Oddly, the more we require chaos, the more we desire order, a seemingly suicidal mass psychology (Lemmings?)

As proponents and practitioners of these much more highly advanced management models based on living systems, I am deeply impressed by the phenomenon of people at numerous levels professing the need for change (a varied move towards chaos) yet cling to order, even as they acknowledge it is a fatal mindset.

In a one hour talk that I give, I spend 5-7 minutes on chaos and order, making it clear that chaos is coming no matter what we want to believe and hang on to, but we have a choice as to how we manage it at the personal and organizational level. The nodding sea of heads is neat to see. Yet, in subsequent follow up, the cling to order is frustrating.

I tie this all into the critical need of humanity and society to rapidly adapt these more advanced models of organization, and again, there is wide level of agreeing to this need to adapt, but in the end, almost everyone thinks it's up to others to adapt, and not them.

How do other reading here, handle this resistance to change? Given your level of experience, +Dave Gray , your thoughts here would be of great value.
 
+Gregory Esau Wow, a lot to unpack here. Like you, I have come to the conclusion that the future of organizations lies in understanding them as living systems, which indeed they are. After all companies are made of people. So in my view the person is the primary unit of analysis in the system. This can include employees, suppliers, regulators and so on. The primary operators are people.

We have become accustomed to looking at the people who make up our organizations in the same way we look at the elements of a machine, applying, for example, the principles of mass production such as assembly lines and standardized, interchangeable parts. But while you may be able to standardize a job description, as you add up all these interchangeable parts you deprive yourself of the very things that people are good at: creative problem solving, ingenuity, learning and so on. So by demanding standardization you end up standardizing on lowest common denominators instead of the kind of peak performance people are capable of.

I agree that we need to rethink the way we design, build and structure our organizations, and that living systems offer the best potential models. I'm not sure if you make a distinction between chaos and complexity, but my focus has been not on chaos but on complex adaptive systems, also known as CAS. The interesting thing about CAS is that they do exhibit a kind of order, although because of their nature their behavior is not predictable. So to some a CAS might look like chaos, because it can be difficult to discern the underlying patterns.

The reason CAS are not predictable is that they contain agents who have a will of their own as well as memory, so just because an agent acted one way in situation 1 does not mean they will act in the same way in situation 2, even if situation 2 is identical. So CAS are learning systems that evolve over time.

Uncertainty, unpredictability and ambiguity are characteristics of CAS. While there are many examples of CAS, the one I am most fascinated by is the city, for several reasons:

1. Cities are CAS that are entirely constructed by human beings.
2. They are demonstrably more productive than companies.
3. We have enough of them that we can make comparisons of different cities, to try to understand why some are more successful than others. We can compare across dimensions like population density, corruption, democratic vs dictatorship and so on.
4. We have about 5,000 years of history that we can look at for examples to learn from.
5. Cities are a mix of intentional design and emergent phenomena, and the management of cities as they grow is very much the management of living, uncontrollable systems.

That is, although a city cannot be controlled, it can be managed. And since "managing what you can't control" is, I think, the key to unleashing the next phase of organizational improvement, I think cities and urban design are great places to look for ideas.
 
+Ed Campbell The thing that I have in mind isn't the everyday work of those workers. I have faith in good workers and believe that they are very important and need yo be treated we'll, however when you look deeper into the future, and think of an overall picture, you would think of the consequences of the "little jobs and retraining". Short term the concept is great, however, over time will become expensive, and unless there is a demand for the product, you will either have a quickly shrinking bottom line, or as a worst case scenario, a closed company. Which is worse in the end community wide, not necessarily just for the company.
 
I agree that we need to rethink the way we design, build and structure our organizations, and that living systems offer the best potential models. I'm not sure if you make a distinction between chaos and complexity, but my focus has been not on chaos but on complex adaptive systems, also known as CAS. The interesting thing about CAS is that they do exhibit a kind of order, although because of their nature their behavior is not predictable. So to some a CAS might look like chaos, because it can be difficult to discern the underlying patterns.

It's down to splitting pretty fine hairs, and semantics at this level, +Dave Gray . I am in 100% agreement that living systems, and complex adaptive systems offer us the best models and ways forward. I know you're pretty busy right now, so if the opportunity comes up in the right context in future posts, I'll tease out the nuances.
Add a comment...