Now you know what Flow is about, let us have a closer look at Lean’s 4th fundamental principle named Pull. Actually this blog’s title better would sound ‘let pull the Flow through your organisation’ ; as Pull means indeed that your operations should Flow according to your customer’s rate of demand. According to Pull, you should only process what has been requested (e.g. ordered) by your customers.
Pull & Flow: Lean’s Yin and Yang?
Needless to mention that Pull and Flow are going hand in hand and that both aim at the same main objective: business efficiency and agility. Like the Yin and the Yang, they are synergic and inseparable. An efficient Flow through your organisation is actually the consequence or the result of Pull: 
there is no Flow when there is no Pull by customers. While Flow focuses on removing any obstacle to minimise throughput times (e.g. the overall lead-time between a customer order and its respective delivery), Pull banishes overproduction. Or put another way: thanks to Flow, you eliminate time related wastes, while Pull helps to fight more quantitative wastes like inventories, work-in-process, etc.
Pull for all organisations?
The answer is rather “yes, but…”. The degree of affirmation, and thus the degree of Pull obviously depends on the nature of your business, particularly its activities and the market in which your organisation operates. Which on their turn determine the process types.
If you deliver very customer-specific – say customized – products or services, like make-to-order, build-to-order or engineer-to-order, then the answer is fully “yes” to Pull. E.g. most architectural firms will not start delivering plans and drawings, resulting from their design and study work, as long as there is no order. Hence, this is 100% Pull.
On the other hand, if you deliver mass products like Fast Moving Consumer Goods (FMCG), or mass services – e.g. passenger rail services -, then you will obviously not wait until a customer orders before processing the good or service. In such a case, you will need to forecast the demand to plan and to optimise your processes. But even then, you will be better off applying Pull principles. In this case, the degree of Pull will be lower, but should not be absent! Notwithstanding you will need to foresee one or more decoupling point(s) somehow, applying Pull will force you to minimise inventories and thus to optimise performance. Following sub paragraphs illustrate this more in details.
Pull and Process types
As illustrated in below figure, Slack, Chamber & Johnston distinguish following process types, considering on the one axis the variety and volume of the product or the service processed, on the other axis the process continuity and diversity:
Project processes and professional services are pulled by nature, given their high specificity, or even their uniqueness. While continuous processes and mass services are characterised by a very low – if any – customisation; and these are thus far from being pulled by nature. Needless to mention that it is for those mass processes that applying Pull principles will be the most beneficial; or other said, where Pull offers the most room for improvement. Bicheno & Holweg explicitly illustrate the relation between this kind of process types and the push-pull boundary:
And this leads us to next topic: decoupling points within a process or a supply chain.
Decoupling Points and Pull
What is a Decoupling Point ?
A Decoupling Point (DP) indicates how far a customer order pervades in an organization, a process, or a supply chain. Stated another way, it indicates where to foresee an inventory to satisfy the customer requests. Indeed, throughput time(s) will depend on how many inventories there are, and their respective location(s) in the process or supply chain.
An FMCG, like tooth paste, available in a (physical) retail shop – which is by definition a local inventory of an end product – enables a quick supply for the customer. S/he can take it from the shelve to respond his/her need. No need to mention that keeping large amounts of end products in local inventories may be very costly when there are many points of sales; e.g. keeping 10 pieces of a product in a retail chain having 500 shops means 5.000 items to keep in stock. Moreover, when this products has variants – like different sizes or colors -, this number of items to be kept in stock may explosively increase; leading to very high inventory costs.
On the other hand, if the product is “out of stock”, the shop does not only misses sales opportunities, but also may get image damage, cause customers expect to get it from the shelve but cannot.
You undoubtedly understand how unrealistic it would be for car companies to be able to deliver their cars in every possible combination – with the many options – at the time of the purchase decision by customers.
The importance of the location of a Decoupling Point
An important aspect of a Decoupling Point’s location is the fact that activities downstream (i.e. to the right of) the DP, are Pull driven, while the activities upstream (i.e. to the left of) the DP are inventory driven, thus planning and forecast driven.
You will understand that the most downstream the DP is in the process or in the supply chain (e.g. DP1 in below illustration), the higher the responsiveness to customers, i.e. the shorter the delivery time will be for the customer. Like this is the case for FMCG’s which are available from local inventories (e.g. in a retail shop). But, however, the higher the inventory costs will be. While the more upstream the DP is located (e.g. DP5 in the diagram), the higher the delivery time will be; though the less costly the inventory usually is. Indeed, not only an inventory of raw materials is normally cheaper than an inventory of more valuable end products; but you will also need larger numbers of stock items. Especially when you have diversified products, i.e. products having many variants.
We can conclude that for goods which are rather cheap (to keep in stock), for which the customer expects (nearly) immediate availability and of which the sales volume is rather easy to be forecasted, the DP should be as downwards as possible, e.g. DP1. For more expensive products, or those for which the market does not expect fast deliveries, or for which the sales volume is more difficult to forecast, you better consider the DP more upstream, e.g. DP5.
How many Decoupling Points?
Given the tension between the cost of inventory on one hand, and their beneficial effect on shortening delivery time on the other hand, determining the optimal number of decoupling points and respective inventories may be a challenging balancing act. The more steps or the more complex a process or a supply chain is, the more DPs you will need to foresee.
Referring to above paragraphs, notice that an organisation dealing with build-to-order or engineer-to-order processes may have no DP at all. Indeed, the only inventory which an architectural firm will have is paper and ink for producing their drawings; which are meanwhile most often replaced by electronic versions, by the way.
Some Pull effects
Kanban, JIT
Though I already mentioned Kanban and JIT (Just-in-Time) as results from the Flow principle, these are also the consequence of Pull, indeed. Especially Kanban, of which one of the main objective is to limit the “work-in-process”. Organisations applying Kanban will surely never process things that are not yet ordered somehow. And the same can be said of JIT: it aims at delivering at the planned – possibly promised – delivery date, and it will – that way – Pull the upstream flow (or process steps) through the organisation.
Vendor-Managed Inventory (VMI)
Where Kanban is a nice example of Pull at the process or at the organisational level, Vendor-Managed Inventory is the example of Pull in the supply chain.
Why is VMI a Pull-based technique?
Imagine a supplier of a FMCG, e.g. tooth paste which is sold in large supermarket chains. VMI means that the retail chain informs the supplier as accurately as possible about the items sold. This way, the supplier knows quickly – ideally in (nearly) real-time – the amount of its product available in the many points of sales.
This really helps to limit the bullwhip effect, which is explained in this previous blog; and it thus contributes to an improvement of the overall supply chain, by limiting inventories.
How IoT (Internet of Things) may will help to Pull
Reasoning one step beyond VMI, the Internet of Things will enable you to trace items and inventories which you want to follow for optimisation purposes, say to improve the Pull effect. In a supply chain, the Internet of Things will be particularly interesting for items that are not registered like at a POS (point-of-sales) system; in contrary with the example of tooth paste in supermarkets, here above.
Imagine you are a supplier of bricks who wants to monitor the current volumes of pallets still available at the numerous construction sites of clients. Putting a sensor on each pallet, that will tell you when the masons start a new pallet – for instance when they remove the plastic package of the latest pallet – could certainly help you to follow in (nearly) real-time the need to deliver new pallets. And hence, it would help you to avoid unnecessarily large stocks of bricks at the many construction sites, which moreover run the risk of being stolen. In which case you may even trace through a geo-sensor that a pallet has been moved away from the site, and where it has been moved to.
How did you apply Pull in your organisation so far? Share your experience through the ‘Comments’ box here below and receive the most complete e-book (pdf of +/-700 pages) on Operational Management, with highly valuable insights, teaching materials and examples which will certainly help you to further improve your operational activities.
P.S.: Please share this information with your Facebook friends and fans, LinkedIn contacts, Twitter followers and Google+ circles, through the share buttons below. Thank you!