A typical situation I have seen in many companies is how the Push system causes delivery failures and poor results because the resources work on other things than the tasks that are in demand here and now. That is, to satisfy the customer.


Example of bottleneck management

One company described itself as order driven. This means that they believed they were working according to a Pull system where the customer order controlled the daily tasks throughout the company. In practice, they worked according to a Push system where resources were managed according to what tasks they had in each department, and what was resource optimal at any given time. The result of this was low productivity, poor delivery reliability, long lead times and poor financial results. Their way of working was that they received an order, which was processed by creating a 3D construction in their computer system, and this was then sent to the warehouse which was the start of their production line. In an ideal world, this could have worked if they had full flow and 100 percent balanced production line. In practice, the line was by no means in balance. The warehouse had good capacity, and they started all orders immediately, but since production had a bottleneck a little further into the process, semi-finished products with ready-made tasks piled up. Lack of space meant led to semi-finished products that had to be stored outside, they snowed down and had to be shoveled back when the production was ready. At the same time, it was seen that the bottleneck machine produced efficiently perhaps 4 – 5 hours a day due to long changeover times.

The simple solution they came up with was to put the bottleneck machine in focus. When an order was finished in the bottleneck machine, a new one was ready: materials, tools and consumables were placed on a rolling table, and it took almost no time to finish an order and start a new one. The warehouse stopped preparing new orders until the bottleneck machine was ready to receive, and the warehouse personnel prioritized preparing orders on the rolling tables for the next job in the bottleneck machine in front of preparing jobs that would not be produced until next week or later.

What was done was to introduce the Pull principle from the bottleneck machine and backwards in the production process and in this way achieve flow.

The result was better flow in production, less need for space when semi-finished products were not produced too early, better quality because storage and many transports led to damage to the materials and parts that were lost. Less need for resources and shorter lead times led to better delivery reliability and lower costs.


Example from Central Warehouse in Functional Enterprise

For a time, it was popular to create central warehouses that were controlled digitally. This was believed to be very efficient in that one knew where all goods were, one could ensure that all available resources could be immediately added tasks by extracting orders with materials or semi-finished products from the central warehouse. In practice, this resulted in a production situation with a total lack of flow.

In a mechanical company, they had such a central warehouse, and they had a computer-based basis for all processes. Between each operation, the order went into storage and then was taken out again when the next operation was to be performed, according to the planning system. The lead time of these processes was very long and eventually they changed their entire production. The central warehouse was discontinued, and the workshop was organized according to a PFA (Production Flow Analysis, ref. John L. Burbidge), to get flow in production based on Pull, instead of the Push system where the central warehouse controlled the tasks in the machines


Example of the Pull principle in a forecast-driven business

An enterprise produced knives. The knives consisted of forged blades, shafts, and sheaths. Procurement of raw materials was based on forecasting and production was mainly needlework and operations performed in conventional machines.

Large batches of shafts, blades and sheaths were produced, which went to storage and were then taken out again for assembly orders in batches of finished knives. These went to finished goods warehouses where knives and sheaths were packed and prepared for shipment. The idea behind this was that production in large batches was more efficient because there was no change over time, but instead it meant stocks of semi-finished products and finished goods in stock, while products were constantly lacking when customer ordered.

They changed production. After all, the transition times were very short. First, they moved the machines together in the workshop for shaft production. The goal was to put the machine so close to each other that a person could produce the shaft from raw material to finished shaft in one operation by moving the product from machine to machine standing almost at the same spot. After carrying this out, they found that they could install a small robot, thus getting an axle production that produced continuously and only stopped in case of tool breakage or other failures, in line with Toyota's Jidoka, or "Automation with human touch".

In the warehouse, a Kanban system was introduced by defining racks for all knife types in the sales warehouse with defined minimum inventory. When the bearing went below the defined minimum, one signal went to shaft production and one to sheath production and when the shafts were produced, they went to the knife assembly. The finished knives and sheaths arrived at the warehouse at the same time where they were packed and placed in the rack. Forging of knife blades was also produced according to the Pull principle, but here the batches were larger.

They introduced a Pull system where they never ran out of knives. The lead time for a batch of knives decreased to 2 – 3 days, and they had no intermediate stock of semi-finished products.


Flexibility vs. mass production

In mass-producing enterprises, flow has been achieved in parts of production with the help of production lines where the progress of each operation is controlled by the pace of the line. This is a way of working that has been referred to since the Industrial Revolution and was used in early production models such as car manufacturing at Ford.

It was said that when Toyota studied Ford's production, they decided to place more emphasis on flexibility than on standard mass production. Scientific Methods by Friedrich W Taylor and Henry Ford were not removed, but further developed to introduce a greater degree of flexibility.

Toyota's adaptation to Ford's production model was developed into what we know as Just in time (JIT) which is a Pull system in production. Materials and tasks should arrive and be performed when the need arises. This is Pull's philosophy in practice and became fundamental to Toyota's broader production philosophy. A way of working that today is often referred to as Lean Manufacturing.

Further development was the focus on flow, which required short turnaround times and small batches, leading up to “one-piece production”. Transition times between operations were reduced or removed. First, by the SMED method (Single-Minute Exchange of Die) which involved reducing the changeover time to less than ten minutes, and then to OTED (One Touch Enchange of Die). The driving force for this focus was to move away from batch production. Batch Production was considered the biggest obstacle to flow in processes. When you produce a batch of 100 units, there will always be ninety-nine units to wait, while it is produced on one unit. This was described as wasteful.


To implement Pull in production

In many functional companies we often have poor material flow and inefficient processes. Attempts are made to optimize machines or other resources and the materials are waiting in large stacks in storage or in production in anticipation of being processed. To achieve flow in such a company, the best principle is to reorganize the resources so that they are assigned to the delivery processes, but this can present problems if, for example, you have heavy machinery that all the processes must go through, or you have many products and processes that depend on the same resources.

  1. The Pull effect in production comes from customer needs. This is what should always govern the implementation of tasks
  2. Organizing in lines is a good approach to flow in processes. Here, the Pull principle is introduced by the line producing if the need for the unit that the line will produce has not exceeded. You can then adjust the speed of the line, or switch to another product when the ceiling is reached.
  3. Prioritize buying several cheap machines, even secondhand machines, and place them where you need dem in the process, instead of one single expensive machine that should serve many processes, although this one is more efficient.
  4. You can organize into cells and ensuring that each product cell has the resources needed to produce its products. Each cell can then act as an autonomous team with responsibility for their deliveries. Of course, there is nothing to prevent a cell from also having the opportunity to use common resources together with other cells. Pull principle is safeguarded by order production in the cell, so that delivery orders are produced according to customers’ demand. If you produce for storage by having defined the level for an item to be produced, this can still be a Pull system.
  5. You can do a spaghetti analyzes where you draw up the way the material moves in production. Start with the first operation, such as inventory, draw a line to the place where the next operation is performed, and continue with it until the process is finished. You can then see how far the materials travel, you can measure the time of each process, and you can enter the operation time, transport time and storage time. Often the transport takes a long time because it involves waiting. When a batch is finished in an operation, it takes some time before the transport arrives and pick it up, and when it has arrived at the next operation, it can take a lot of time before those who will perform the next operation are ready to start. Often one gets aha experiences with drawing up such a diagram and is motivated to make a line or cell organization, in whole or in part. Of course, the spaghetti diagram has the same relevance in an administrative process too.


Pull term is an accepted principle in production.
There are still many voices against the term Lean because this is perceived as predatory practices, redundancies, mergers into large units and poorer working conditions. Perhaps Pull would have been a better term to front this way of working?