Demand Planning in Uncertain Times

Abstract

Demand planning is just not what it used to be…

Article

Introduction — This Isn’t History Repeating Itself

The pattern of life has been so radically altered that forecast patterns of the past may not work anymore. (You don’t have to read anymore if you get that statement!) Actually, my contention for decades is that using history for your forecast is not always that useful, anyway.

Did you ever notice how ‘out of the blue’ some new product, company or idea comes along and upsets the apple cart of the product sector? When you talk to the founder or inventors, turns out generally they invented such product because they could not find what they were looking for themselves. The only difference between them and me, is they had the smarts and stamina to actually invent the thing that we all needed, and the gumption to get it funded to build it. and then the perseverance to get it sold into the marketplace. Doesn’t that make you wonder about all those latent unfulfilled desires?

But right now, that might not be on your mind. Right now, supply chain practitioners are dealing with multiple years of economic up and down turns; major societal shifts in the needs of a multi-generation and global market; geo-weather disasters; and now pandemics. Right now, end-to-end demand has been a big surprise even to the makers of these products, and thus fulfillment of demand has been a huge challenge for many market sectors. And I mean end-to-end — from food to toilet paper, from the kitchen blender to the plumbing.

Food sources have become undependable, just as the channel to buying food and what people want to buy has changed. We shifted from a going out to eat society to home baker, ensuring that stocks like whole wheat flour, yeast and baking powder run out as customers wait for weeks for restocks.

The now stay-at-homes without service repair workers coming to the house has turned us from a procurer of services to the home improvement stars. I recently read in Consumer Reports that Americans have been upgrading their kitchen sinks, purchasing the latest kitchen gadgets and even buying high tech toilets, and many manufacturers have run out of fancy models, especially bidets!

Lawn and gardening products — sold out! Some of the biggest national suppliers of plants and flowers apologizing to their customers in May that they were not shipping anymore for the season. They were already sold out for spring and summer and only planning for fall shipments.

How about some pajamas for the work-at-homes? Sold out. And then there are all those gyms. Can’t go to the gym? Build your own home gym with a range of products from simple inexpensive exercise bands and some dumbbells all the way up to fancy Peloton subscriptions with high tech bikes or treadmills. Or run in your neighborhood, avoiding the few passerbys you may encounter, with a spike in demand of the best shoes to keep you bouncing along. And keeping the kids content? Home crafts, trampolines and above ground swimming pools are in hot demand.

The lists go on, and I am sure you can supply from your own life and your company how requirements have changed. And that is a clue for the quiz question coming soon.

Establishing a New Foundation for Forecasting

So, if we can’t rely on history, what do we rely on for demand forecasting?

In one word: Knowledge. For decades we have been ranting and chanting about customer centricity, yet most of the world products are, well, product driven. So, now it is time to really get to know the customer. And they are leaving clues about themselves everywhere.

That is where AI and Machine Learning come in. With ML I can access huge amounts of data from many many sources and pull that data together in ways that allow me to see the interrelationship between groups of consumers, their habits, income, buying patterns and even their chitter chatter to extract some latent demand.1

Data from sources like Pinterest, Yelp, Google, YouTube, Facebook, Twitter, consumer or market research data services, and specific sites that cater to the relevant category2 are all sources of data to observe consumers. This coupled with your customer data bases can create a gold mine of foundational knowledge of which to then begin to build hypothesis models.

Techniques like neural networks and deep learning have properties of associating and discovering attributes that may not be obvious without deep searching. For example, in deep learning, a program which processes web data over and over again can associate data in many different scenarios or context. This can then be coupled with advanced forecasting algorithms that can look at the data and extract new or changing patterns. And those changes are key.

In traditional forecasting we are looking for ‘absolute numbers’ — how many widgets per week. But in volatile markets we want to look at the curve: how is it changing (up or down), how fast is it changing and how much is it changing. This takes a different kind of math.

Time to Step Up!

Supply chain planners are an above average IQ kind of group. That is why we have been calling them Supply Chain Scientists. And why not? Their skills demonstrate a systemic, analytic and mathematical capability. And this is coupled with the discovery of new knowledge. Sounds like the ‘process’ of science.

Time to step up and adopt some new methods, if you haven’t already begun. Yes. There is a lot to learn and organize here. But surely it’s worth it — like survival.

In years past many advocated for manufacturers/brand companies to leap over the channel to gain knowledge about their end customers. The web held that promise. Many did not choose this path. But now, even the channels you may have relied on in the past for market insight and demand are kind of broken. Or maybe even they are gone.

So, for the sake of survival, step up.

__ __ __ __ __ __ __

Further Reading:

Artificial Intelligence / Machine Learning Collection

My Demand Manifesto

_____________________________________________________________

1 For more on AI/ML read: AI in Supply Chain — Some DefinitionsReturn to article text above
2 For example, if your products are ingredients, flours and grains, there are lots of cooking channels, publications and hundreds of bakers’ blogs. Or if your products are sports equipment, we have the relevant publications and broadcast channels where sports enthusiasts congregate, and so on. — Return to article text above


To view other articles from this issue of the brief, click here.

Scroll to Top