Splitting Hairs?

“Just definitions either prevent or put an end to a dispute.” – Nathaniel Emmons

Lora Cecere, a well-known industry consultant and author of the Supply Chain Shaman blog recently made the following prediction about :

“Distribution-centric industries are not all pull. The flows are a combination of push and pull based on demand shaping programs. As a result, flowcasting may be a good tool for turn-based volume, but lacks the depth of analytics to help with promoted goods and new product launch.”

To be fair, she was actually speaking about 's Flowcasting software (which got its moniker in the RedPrairie merger), but based on the wording, one could also attribute her comment to the Flowcasting idea.

Whether or not you agree or disagree with her sentiment boils down to how you define the following:

  • Distribution-centric
  • Push and pull
  • Demand
  • Promoted goods
  • New product launch

Ask any 10 supply chain people to define these terms and you're likely to get at least 6 different answers. While Lora is an experienced and respected supply chain expert in the manufacturing space, I've spent my entire career in . I have often found that I speak a different dialect from my upstream supply chain brethren.

How you define each of these terms is critical to determining your view about what a supply chain does, why it exists and how it operates.


Each successive step in the supply chain (manufacturing, distribution, transportation, merchandising) adds cost to a product and increases its value to the consumer. However, a key tenet of Flowcasting is that value is not fully realized in the supply chain until the product reaches the point of final consumption. In that regard, any company that manufactures, distributes, hauls or sells product that is meant to be purchased by a consumer should not consider themselves as part of a “distribution-centric industry”. Distributing the product is only one part of the – if products are not purchased by the end consumer, then everyone – from the manufacturer to the retailer – will be out of business in short order. By my definition, all of the players should consider themselves part of a “consumption-centric” industry.

Push and Pull

By a wide margin, these are the two most overused words in the supply chain lexicon. We assume that everyone has the same definition, but if consensus has been reached on what these words actually mean, I have yet to hear it.

My definition of “push and pull” is an extension from my definition of “retail supply chain”. The way I look at it, the supply chain hasn't done its job until the customer has made the purchase. In that context, the notion of a “push” supply chain sounds a bit absurd. I imagine a customer wheelling a shopping cart through a retail store while the staff fill the cart – against the customer's will – with product that she doesn't want. The point is that if you believe the supply chain ends with a customer purchase (and you also believe that the customer must choose to buy it), then the idea of “pushing” product is really a shell game – in the end, the supply chain can only be driven by a “pull”. In the case where supply exceeds demand (to be discussed later), the only option to get the product completely out of the supply chain is to “encourage a pull” from the consumer through markdowns and clearance pricing.

In the case where supply is constrained, decisions need to be made as to where to locate the product where profit will be maximized. This can mean forward positioning the product to DCs and allowing the stores that are selling it quickly to “pull” it the rest of the way. Flowcasting fully accounts for the constrained supply scenario and provides the most up-to-date information possible to properly ration when this occurs (i.e. when a customer facing stocking point has a stockout in its future, the “pull” is disabled until such time as the location is back , which may not ever happen for short lifecycle products).

The point here is that there's more to managing a supply chain than just replenishing the and if the whole plan is not based on expected sales at the store shelf, then it will not be possible to do forward looking capacity , and joint business planning with a single set of numbers, all of which are necessary to maximize sales while minimizing supply chain resource usage.


For any enterprise that participates in making, distributing or merchandising product that will eventually find its way into a consumer's hands, “demand” (or more specifically “independent demand”) must mean expected sales at the store shelf. When a manufacturer considers demand to be “customer order demand on the DCs or plants”, they are talking about dependent demand which does not exist without expected sales at the shelf. In the Flowcasting philosophy, managing dependent demand is a waste of time and a huge driver of inefficiency in the supply chain. As Dr. Joseph Orlicky (the father of Manufacturing Resource Planning) once quipped: “Never forecast what you can calculate.

Promoted Goods

I'll make a distinction here between “regularly selling goods on promotion” and “promoted goods”.

Regularly selling goods on promotion are easily planned in Flowcasting through an update to the forecast to account for the expected lift in sales during promoted weeks. The increased forecast “pulls” product to the correct locations just in time based on inventory levels, rounding rules and all the normal stuff that drives the supply chain for these same items during non promoted periods (with the exception that safety stock may increase ahead of the promotion to execute display setups).

In contrast, “promoted goods” (by my definition, not necessarily Lora's) are items that are limited-run, short lifecycle items (perhaps with special displays) that are not planogrammed in the store. They are placed on the sales floor, sell out and are discontinued. Again, I fail to see how this is different from the short lifecycle product scenario. Before 10,000 units of a promoted product is made or bought, there needs to be a plan to sell it. Where will it sell? When will it sell? How did we arrive at 10,000 units? At what price will we sell it to make it a profitable venture but still not have any carryover stock? These are all “pull” questions. If they can be answered before commitments are made to make or purchase the product – and they certainly should be – then there's no reason why a promoted item cannot be planned as a “pull” for all of the reasons mentioned previously.

New Product Launch

Flowcasting is the only process that can plan the expected sales of a new product discretely and independently (with regard to both magnitude and timing) from the “pipeline fill” requirements to ensure that both merchandising and safety stock needs are met and sales are supported with the correct timing for both.

If a product is a “new variation on an existing product”, then a forecast can easily be created by copying and scaling existing sales history. If it's something that the world has never seen before and is completely unlike anything else, then additional guesswork and is needed to produce a forecast. In either case, the assumptions driving the sales forecast are documented and executed within the Flowcasting process.

When sales materialize in an unexpected way (i.e. higher or lower than expected for any selling location), Flowcasting automatically replans the supply chain and alerts people to the mismatch on the day it begins to happen.

Short of an actual crystal ball that can predict the future with 100% accuracy all the time, I'm not sure there's a better way to manage a new product launch, at least in a world where retailers and manufacturers are working closely together in a single process, which is a trend that's accelerating.

I'm certainly no shaman myself, but I have only one prediction: A lot more discussion and debate about this over the next few years.

Leave a Reply

Your email address will not be published. Required fields are marked *