menu icon

Pull Flow Systems for All

By Bernard Milian

Pull Flow, an Old Story… 

In 1987, as a young engineer at Philips, I had the chance to take part in a study trip to Japan and visit about ten companies there. In most of them you could see in each workshop card boards. In one of the factories I asked our guide what it was. He was astonished that I asked the question – it was so obvious: it is a Kanban board, to replenish supplies. That is normal. It was a non-event, a practice that was not questionedhad been in place for decades… but almost absent at the time from our western factories. Kanban dates to the ’50s. Drum-buffer-rope dates back to the ’80s. These are old concepts. Are they adapted to our modern, technological age?  

The benefits of pull flow have long since been demonstrated 

Under the impetus of the Lean wave, level loaded pull flow has gradually been adopted in several companies in a variety of industries. However, its adoption has been predominant in the automotive sector, but often timid elsewhere. I have implemented pull-flow tactics throughout my career, in industries as diverse as electronics, furniture manufacturing, metallurgy, medical devices, and so on.  

I no longer need a theoretical demonstration because I have seen it in practice over three decades: pull flow always brings a substantial improvement in terms of availability, stocks, lead time, visibility. Always. Without exception. To test it is to approve it. 

And yet pull flow systems are not widespread in our companies. 

If pull flow is so powerful and notoriously superior to push flow, why is it not common practice in business, as were already kanbans in 1987 during that trip to Japan? Let’s face it: Pull flow is ubiquitous in the automotive industry, but largely absent in other manufacturing sectors. In recent years, I have been involved, for example, in pharmaceutical, medical device and aeronautics companies – whose predominant operating methods are MRP, supported by in-house logic developed in spreadsheets.  

Sometimes there is a bit of pull, but it is limited to a small area: one piece of equipment, one workshop, a few suppliers. There is never an end-to-end pull flow. When we ask the teams about kanban, either it is known by name but not practiced, or “it’s not adapted to our industry”.

The concepts of pacesetter workstation station (constraint, drum), task, queue management (time buffers) in those industries are often ignored. Why, after all these decades, are we still here? 

Implementing pull flow has always been complicated. 

Setting up a pull flow system on the scale of an entire company, or even an extended supply chain, requires an alignment of planets that is still rare: a clear vision of the management team, a shared multidisciplinary will, a sharing of concepts and a common language, and adequate information systems. Companies that have widely deployed pull flow within their plants and distribution networks, such as Valeo, Faurecia and Schneider Electric to name but a few, have developed in-house a body of methodology, training and IT systems that facilitate the deployment and maintenance over time of the pull flow model. Dynamically readjusted kanban loops, visual management, planning through a few control points using sequencers (heijunka), queue management are the norm. If you do not have the strike force of these companies that have been able to develop their own logic, the implementation of pull flow is much more complicated and often seems out of reach. 

Several factors are at work:   

– No training and shared concepts 

Prevalence of flow, sharing of management concepts, transversal understanding between supply chain, production, finance, sales, quality, etc. require training to align teams. Even if Lean concepts have gained ground, most companies do not have this type of cross-functional training, and the smallest project requires a lot of energy and power of conviction. 

– A difficult theoretical demonstration 

How many times in my career have I had to try to demonstrate the efficiency of a pull-flow system and build a business case to get a budget approved? It is very complicated because it often goes against traditional accounting. This old story between flows and costs. I was tired of the war and most of the time I took the piece by proposing an inexpensive pilot to demonstrate first that it works well on a factory scale. When you are talking to a board of directors of a group, the story is different. You need to demonstrate through theory and a solid business case. If you do not have a leader in the management team who is already largely convinced, you’re done. 

– Unsuitable information systems 

In the 21st century, working without an information system is out of question. Digital has invaded our daily lives and is indispensable in business. ERPs do not excel in their ability to facilitate pull flow. They are all wired with an MRP logic and at best sometimes integrate a basic Kanban logic. Reconciling kanbans and the information system remains a challenge.  

– No standard, a multitude of proprietary systems 

There are dynamic electronic kanban models on the market and several “schools of thought”. There are some excellent approaches, no doubt about it. Unfortunately, it is a jungle of proprietary systems: difficult to navigate for non-specialists, and when you adopt the system of this or that software company, you are locked into that model. 

– A necessary adaptation 

Every business is different. The pull flow model must therefore be adapted to this reality. In the automotive industry you make 1000 parts per day on 10 references. In aeronautics 10 per day on 1000 references. Copy and paste does not work. 

Companies need to acquire skills in designing the operating model, i.e. internal team members who will be able to adapt the principles to their specific flows. Mapping the flows, identifying the real constraints, positioning the protections, defining the planning and executions logics. “The method does not suit our industry” is a false excuse. Pull flow works in all industries but must be adapted. It is therefore necessary to develop and maintain these skills, and if need be, to be supported by good consultants from time to time. 

Twice in my career I have moved from an automotive equipment environment to other industries that seemed to be at the opposite end of the spectrum: 

The first time I joined a mattress and box spring manufacturer. Restful? Not at all, much more complicated than the automotive industry: a proliferation of configured items, and promotional operations that multiply sales by a factor of 2 to 5 over a few weeks. It is much more relaxing to feed a car manufacturer’s clocked line! The pull-flow model we set up was therefore quite different, and for promotions used low and high forecasting assumptions, and adjusted to the daily pace of order taking. We would call that today demand sensing, but that was in the 90s… 

The second time was for a single-use medical device company, supplying its products from Asia in sea containers… Yes, yes, pull flow works very well in this context, with a structural lead time of several weeks, but here too it has to be adapted.  

– No link with S&OP and the medium term range 

“We prepare with MRP, we execute in the short term with kanban“. That is what was taught in the 1990s. Yes, but here it is : between the short-term horizon of the kanban and the medium-term horizon controlled by MRP based on safety stock and forecasts, there is no connection, because the two approaches are based on different logics and parameters. Moreover, if pull flow is not adapted to adjust for the S&OP inputs (for example seasonality, Chinese New Year, or a maintenance shutdown), it is doomed to failure. The methodological framework and adapted tools have long been lacking for this.  


Why are DDMRP and the DDOM game changers ? 

The Demand Driven methodology combines good practices from Lean, Theory of Constraints and MRP to establish a pragmatic pull flow model adaptable to any environment, whatever its variability.  

The Demand Driven Institute has been able to develop and “industrialize” this model in just a few years, which greatly facilitates the implementation of end-to-end pull flow. Transformations that used to take several years and require a great deal of effort, if not failure along the way, are now possible in just a few months.  

– Training and shared concepts 

The adaptive business model (DDAE) and the available training, conferences, books, simulation games facilitate the alignment of teams for a common understanding. This is an essential element for a successful and sustainable approach. 

– Easier business case demonstration 

The Demand Driven model can be simulated. The impact on inventory and service can be assessed a priori, with a good degree of confidence. Finally, it is much easier to substantiate a business case. 

– Standard framework 

The Demand Driven Institute has established a methodological standard and compliance criteria for software solutions. DDMRP Compliant, DDS&OP Compliant, DDOM Compliant – you will no longer be prisoner of a vendor, the methodology and model remain valid. 

– Information systems 

Did tell you about Demand Driven Technologies’ solutions  ? 

From Demand Driven S&OP to visual queue management on the shop floor, you have all the features you need to set up an end-to-end digitalized pull flow. Some would call this Lean 4.0.  

– Emphasis on modeling 

The first step in the Demand Driven methodology is the design of the model. The Demand Driven Leader training is more specifically dedicated to developing design skills. Dedicated consultants are available. 

– Linking with S&OP  

The DD S&OP and Adaptive S&OP torque finally allows : 

To ensure consistency between short-term cash flow and medium-term cash flow. 

To put the church back in the centre of the S&OP process: it is not a question of establishing an illusory balance between demand and supply, it is a question of constantly adapting the company to changes! 

DDMRP and more broadly the demand-driven adaptive business model facilitates the implementation of pull flow, and makes it available to all enterprises, large and small, in all sectors of activity. It is simple, fast and efficient. If you are faced with changing demand and difficulties in adapting your operations, do not hesitate to explore the opportunity !

If you are interested in learning more about our Demand Driven Supply Chain Software solutions feel free to contact us and see how DDMRP can improve your supply chain. 

Get in Touch

Share This Story, Choose Your Platform!


Recent Posts

Sign up to our Newsletter

You may also enjoy

A black and white checkered flag is being held up against a clear blue sky with a few scattered clouds.

Requested Date or Promised Date

Service rate – on time in full – is THE most important performance measure for a supply chain – it measures our ability to deliver