Recently read an awesome book, A More Beautiful Question, about inquiring and asking more than telling. Should be required reading for everyone. Got me thinking though…
Most of the advice and thinking about supply chain planning (i.e., demand & supply planning) is about developing and applying new forms of advanced analytics to the wrong question.
My advice: find a better question! Then you’ll realize you should never forecast what you can calculate!
It’s 1946 and a young Lee Iacocca finishes engineering school and joins Ford Motor Company – at the time, one the most prestigious companies in the world. Lee brings countless suggestions and ideas to his boss, only to hear that same question asked. ”Where has that worked before?”
Undaunted, Lee finds out.
Iacocca learned from his boss a fundamental lesson: if an idea is truly innovative, you’ll find examples of its successful implementation scattered throughout history. The secret to successful innovation is to import a tested and reliable idea or concept into a business situation where it has never been used.
Repurpose the proven.
Almost everyone working in supply chain has heard of Distribution Resource Planning (DRP). Did you ever wonder where idea for DRP came from? Well, from repurposing another concept. That’s when, back in the 1970’s, a young Andre Martin learned about the concept of a bill of material – a simple idea to calculate components and parts needed to produce a product.
Andre was struggling with how to better manage the flow of inventories between his distribution centres and factories. When he saw the BOM concept he got to thinking. Couldn’t he flip the concept and use it to calculate a bill of distribution?Sure he could. He repurposed the BOM idea and the rest is history.
A number of colleagues and business associates have complimented us on the Flowcasting process. Many refer to it as simple and elegant. Take a POS forecast, by item, by store and effortlessly transform it, apply simply calculations to it, until it speaks the language of the supply chain.
But the idea is not quite so brilliant and we’ve seen this movie before. Frankly it’s an extension of the idea of Distribution Resource Planning (DRP). The idea has been repurposed to extend the thinking and the logic to the store.
While it’s nice to hear people refer to Flowcasting as a breakthrough innovation and to us as innovators, you and I know the truth.
We’re really practical repurposers.
Have you every noticed that your views on things change when you’re forced to “put your money where your mouth is”? So, you really think that the Leafs will win the Stanley Cup? Wanna bet 500 bucks on that?
Brings you back to reality, doesn’t it? The same concept applies to demand planning/forecasting.
Bias (systemically forecasting too high or too low) is the #1 killer of good forecasts. The biggest problem with bias is its insidious nature – people often introduce bias into their forecasts with the very best of intentions. It may come as a result of being optimistic by nature. Or out of a desire to provide good service, you may feel compelled to add “a little bit extra” to your forecast, just in case.
You know that the demand forecast needs to be a realistic estimate of what you actually expect to happen. Everyone is planning on it. So how can you be sure that you’re not unintentionally killing your forecast with good intentions?
Use the $500 rule. What’s that?
$500 Rule: If you were forced to bet $500 of your own money that your forecast was as close to the truth as possible, would you keep it or change it?
Two Dimensions of Demand Planning
Suppose you were guaranteed that a specific investment stock will increase in value by 50% but you weren’t told whether it will happen in 5 months, 5 years or 5 decades?
How much would this information be worth to you?
Not much, right? Why? Simple – you need to know both the percentage increase (quantity) and the timing (when). Demand planning is very similar. The purpose of the demand planning process is to anticipate how much of a particular product your customers will want and when they will want it. So, to do a proper job of demand planning, you need to consider two equally important dimensions:
The purpose of the demand planning process is to predict BOTH the quantity and timing. You can’t have one without the other.
A slightly different perspective on Big Data, especially how it pertains to demand planning.
Scott Armstrong is pretty knowledgeable about forecasting processes. In fact, he might be the world’s leading authority. Armstrong is a professor at the Wharton School of Business at the University of Pennsylvania. He’s dedicated his whole life (and he’s 74) to studying forecasting and his book, Principles of Forecasting, should be considered a bible to anyone interested in the field.
Us supply chain planners could learn a lot from old Scotty.
For example, one of his gems of wisdom is about complexity of the forecasting process/model. He states, “there’s been no case in history where we’ve had a complex thing with lots of variables and lots of uncertainty, where people have been able to make forecasting models or any complex model work. The more complex you make the forecasting process the worse the forecast gets”.
Hmmm…makes you wonder, eh?
We have entire companies and self-proclaimed experts who are advising companies about how to factor in Big Data, the weather, and a whole host of other variables, all with complex models aimed at sensing and translating and responding.
Maybe most of these folks are fools? After all, they are forecasting what should be calculated. And the farther away from the consumer you’re forecasting, the more variables and constraints you’ll need to try to factor in. Exactly Armstrong’s point.
Don’t be a fool.
Never forecast what you can calculate!
Dan Gilmore, in his excellent work with Supply Chain Digest, recently presented research that chronicles 40 years of research and measurements on out-of-stocks in the supply chain. Despite $billions spend during this time, there has been NO progress. Here’s a respose to this research from a pioneer in supply chain planning, someone who just might have the answer…
Dan,
The Out-of-Stocks problem in the CPG Industry has been around a very long time and will not go away until we completely change our thinking on how to manage the end to end retail supply chain.
First, we need to acknowledge some basic truths learned the hard way:
Today, it is possible for a retailer and a manufacturer to model and connect their retail supply chain inside a single computer and drive it, end to end, from a unique store level sales forecast.
Today, it is possible to resynchronize every node making up a retail supply chain every day based on changes in store level consumer demand and thus completely eliminate the “Forrester Effect” (also called the Bullwhip Effect).
Today, it is possible for a retailer and a manufacturer to use this model of their retail supply chain and manage it as if one company were managing it.
Today, it is possible for a retailer and a manufacturer, using this “business model”, to create an Integrated Business Plan, identify gaps and take corrective action while there is still time.
Today, there are no secrets anymore on how to manage the end to end retail supply chain, from the store to the factory, and significantly lower store out-of-stocks as well as reduce retailer’s and supplier’s inventories and operating cost at the same time.
Today, it is possible to achieve a 99.8% On Shelf Availability (OSA) at the store while reducing overall supply chain inventory buffers at the same time.
Today, there is a proven and well documented business process, and enabling technology, to actualise all of the above and it even has a name: It’s called “Flowcasting the Retail Supply Chain”
André Martin
A great report on recent trends and directions in retailer – supplier collaboration. You’ll notice that Store-Level DRP (the technical/system name for Flowcasting) is acknowledged as an enabler for improved collaboration and performance. To view the report, click here.
Collaboration promises much to the retail supply chain, and rightly so. So why have results to date been spotty at best? It’s largely as a result of a lack of vision regarding the retail supply chain and not applying common sense to the collaborative process.
Retailers and their trading partners are beginning to understand they are not alone. The retail supply chain does not act as a series of islands – each independent entity working for its own purpose. Rather, smart companies are beginning to understand that they are really working as part of one, completely integrated network that is designed (or should be) to deliver products to their end customers.
To enable that view requires new planning processes and enabling technologies. In simple terms, the planning processes of all players in the retail supply chain need to be connected – much like the network is connected – to allow products to be planned to flow from production to consumption. The result? Unprecedented visibility where all partners in the value chain can see all future network flows – for a long planning horizon (52 or more weeks say). Visibility is the driver of productivity and top and bottom line improvements.
Visibility also enables improved collaboration. Since all partners in the value chain can see what’s expected to happen, collaboration is put on steroids. It’s a result of the information available – the quality, not the quantity.
In fact, I would argue that collaboration will become more important – there will just be less of it. No need to collaborate on the plethora of retail forecasts and planned orders, since these will be automatically translated into the requirements and product flows of all other trading partners. Since everyone can see what’s planned the new mantra of collaboration is “silence is approval” – as a trading partner, unless you indicate otherwise, your ability to deliver is assumed.
Smart companies will collaborate only where they believe it is necessary and worthy of each partners time. That could be on promotional forecasts, plans for new items, ideas and concepts about product flows, supply issues – in fact, collaboration will become the exception, not the norm.
Isn’t that kinda how a good marriage works? You certainly don’t collaborate on every decision you make. Otherwise, you’d go crazy and there would be no impact on the resulting decisions. No, you likely only collaborate on important things, where each of your input is required. And come to a decision you’re both comfortable with.
It makes sense in life and, as it turns out, in supply chain planning too.
At the NRF conference, Kraft and Sam’s Club talked, along with our co-author Andre Martin, about how they are using Flowcasting to improve sales and profits for both organizations. Here’s a snippet from the Big Show’s blog
http://blog.nrf.com/2013/01/13/kraft-sams-club-connect-the-consumer-to-the-factory/