But, face it. All companies have founders and, often, the whole family jumps into the business. From Amazon to Apple to Intel; from LL Bean to IBM; Walmart, Sony, Samsung, and Macy’s to Forever 21, on and on, the founders dominated, often until their deaths and often left the business to the next generation.
As for Forever 21, it is a familiar story.
“Forever 21’s missteps, combined with industrywide changes in consumer tastes and shopping habits, will have far-reaching effects for thousands of people who work for the company, its vendors and malls.”2
Summarizing their woes here:
- Lack of understanding of local markets and buying trends
- Lack of insight into changing customer tastes, values, and buying habits
- Selecting poor store locations and formats
- And most of all, merchandise planning and forecasting errors.
“The business was also making errors with the sprawling store base. Merchandising was based on the previous year’s sales, and Forever 21 bought too little inventory in 2017, then too much in 2018, the filing said. It also duplicated merchandise by designing for ‘styles’ like weekend or work looks, rather than categories like tops or dresses.”3
And they are not alone in this lack of insight and execution on demand. Recently, Thomas Cook,4 a 178-year old company, and a familiar sight to anyone who has ever been in an airport, filed for bankruptcy. They, too, suffered from lack of insight into changing customer requirements. The shoppers have changed and the buying habits have changed. It seems that although Thomas Cook is in a different industry, they are plagued by many of the same problems retailers have.
Changing demand plagues every industry. Oil and gas5 as well as coal6 are other categories seeing layoffs continue, and old equipment will not be replaced. As producers struggle or go bankrupt, all the industrial manufacturing, transport, and service companies that support them also struggle. There are cascading effects experienced in every industry. It is not only the enterprise that is affected, but employees who lose jobs, neighborhoods that decline,7 and customers who may depend on or prefer the goods and services of that company.
So why couldn’t they all see it coming?
Yay! We got the order! Who hasn’t been excited when they close that big deal? But, if you are a supplier to a company, it behooves you to not just rely on that order. It behooves you to look at the big picture of the economy, markets and the end-customer, and the position of your specific customers in those markets, not just at the PO. Suppliers often go bankrupt along with their customers.
If you are a supplier, you need to have deeper participation — collaboration — in merchandise planning, forecasting, and managing the flow of product into those channels (along with the cash flow into your pocket) as you expend on material, labor, and capital to meet that demand.
If I were a supplier to the myriad of retailers who are shutting down or downsizing, I would be pretty distraught right now. Or maybe worse — also having to shut down due to lack of cash or piles of unsold merchandise.
So why don’t trading partners demand more? They should insist on participating in market studies, forecasting, and getting automated data from forecasts, POS, and daily inventory status, at the minimum. And this goes for logistics/transportation partners who need those forecasts in order to plan their capacity, labor, and schedules.
Solutions Are at Hand, but We Don’t Use Them
Frankly, there is no excuse.
How long did it take to develop that spreadsheet (when, today, a product can be bought on the web and installed in a few hours)? That time could have been better spent making sure you got good data from reliable sources in your company to feed into your “system.”
The range of demand-planning products is broad and plentiful. No doubt at the high end they can be pretty pricey, but they also come with profound ROI (and smart, experienced people from the software companies, usually, to help you implement). In the mid-market these tools are pretty sophisticated, too. And, of course, there are those under-$1k products aplenty that really have a lot to offer. A simple forecasting product is a lot less expensive than the two-weeks’ time a typical office worker spends designing a spreadsheet. And now, you are on the hook forever to maintain it. Ouch!
As users get more versatile with spreadsheets, they develop more sheets. The cycle continues — more employee time spent with less integration to meaningful and important data, and the lack of understanding about just exactly what that spreadsheet is doing. That is a big issue — lack of trust in the data, for starters. And when there is employee turnover, who inherits that old work? Can they figure it out and even use it? And more importantly, spreadsheets are just not up to the challenge of absorbing and analyzing modern supply-chain data.8
What Do You Do Now?
The point of reports — data — is to take action. But spreadsheets and unintegrated islands of technology won’t allow that since they don’t propagate into the automated systems or across departments within the organization (what to say of trading partners).9 Demand integration is a real challenge. How will data flow to other critical processes from supply planning, product design, logistics and so on?
A real curiosity is that the higher up we go in the organization, where the most impactful decisions are made, the less automation and integration there is. Spreadsheets, or sneaker net we used to call it, dominate. Though hand delivery is now replaced by emailing spreadsheets, we haven’t progressed very much. In a recent research report by PRG and Supply Chain Management Review, the research showed that only 37% of S&OP users had some kind of automation of S&OP and, of those, only 10% were using an integrated IT system.10
However, it’s not just spreadsheets that are the culprits in poor demand integration. Large ERP companies like SAP and Oracle have many software packages which require the user’s IT departments, big system integrators, and/or process-specific integration platforms to achieve data integration. I was privileged to work with the Marine Corps a few years back and they also had a disparaging term for this lack of integration and automation — “swivel net,” which meant the users had to swivel back and forth between various systems, often retyping data from one into another. It gave me a queasy feeling to consider that our soldiers on the front line were depending on that data, i.e., receiving supplies based on orders from those systems to defend America and our allies.11
Building a Foundation — Innovation Continues
It should be of major concern to any company that while they are relying on past approaches, innovation continues. You get left behind as more enterprising competitors strive for better data, smarter systems, and they invest in their teams with training and support to keep their company vibrant and thriving. There are foundational capabilities that need to be in place. Getting good at — even great at — demand management should be considered a continuous improvement, not a “Hurray, we got the project done” attitude.
Figure 1 shows a foundation for demand management. Not all the capabilities in this model will apply to all types of firms, of course. But the point is that the essentials must be in place before we leap into the new stuff. Additionally, continuous improvement is critical because people only learn so fast and organizations only change so fast.
Figure 1: A Demand Management Model for 2020
Why We Need to Innovate Now!
As markets have become more diverse and volatile, a wider view and newer algorithms are needed.
Innovations to Deploy Now
It is curious to me why companies don’t collect — in any way possible — gobs of customer data. They might start with some manual approaches such as focus groups.14 Today, we have automation. History might be interesting, but looking forward (it’s forecasting, right?) is better.
Consumer Insights: This is where newer approaches like gamification,15 customer sentiment analytics, and so on become really interesting. Customer sentiment16 technology is an AI/insight-creating engine which can use a variety of input data, mostly from the web, but, of course, many other sources that your company should have access to. In fact, AI and other useful analytics have been around for a decade,17 though, of late, they are in a rapid development cycle. Customer insight plus other economic and market condition information should be strong input into planning, whether for current allocations or the next season.18
Tapping into today’s demand: What customers are doing today — right now — is critical. Traditional forecast data takes time to coordinate among the stakeholders. That data should be verified and possibly adjusted due to what is happening in the near-term, e.g., the factory did not get the word that consumers are concerned about X ingredient, and it continues to produce the product. The warehouse fills up, now with a low- or no-demand product.Or conditions in a local area have changed, for example, colder weather has persisted and there is no need to ship the bathing suits. Maybe transfer warmer-weather products from other stores or the warehouse, today.19
The fact is, these kinds of problems are already being solved as progressive demand-planning solution providers have been including these capabilities into software releases.
You can also get started by just getting daily and richer B2B/EDI feeds from your channels, such as POS data, inventory status, and more up-to-date demand (which you should be doing, anyway).20 Supply Chain software companies as well as B2B/EDI firms who have established connections with the major enterprises are a real advantage to tap into.21 They have already done most of the work to help you get that data. Then, there are specific applications such as dynamic allocations, real-time demand/supply matching, and so on for making mid-course corrections, as well as continuous planning in transportation22 that lets you act on the information.
Machine Learning – nother area of contention, but of increasing value within AI, is Machine Learning (ML). Just to demystify a bit here,23 ML is the technology’s ability to automatically learn and improve from experience without being explicitly programmed. ML focuses on the development of computer programs that can access data and use it to learn for themselves. Then recommendations are made. But that does not happen without direct involvement from users — in AI speak, supervised learning.
As you don’t let your teenager drive off in the family car without some training, users train their systems and then let ML begin to parse and analyze, and make recommendations. Just like our kids come home with great ideas, we may — or may not — adopt them.
Even though it can be challenging to get an accurate forecast, demand management is the most knowledge-enriching process an enterprise can do. It tells you about your markets, your customers, your partners. These are essential keys to success. Thus, it behooves firms to continuously commit themselves to innovation and evolution in demand management. As Aristotle said, “It is through knowledge that I gain understanding — and understanding lets me do by choice what others do by constraint of fear.” We can’t wait for fate to befall us; we have to constantly challenge ourselves for excellence. At the risk of sounding like my mother, just because things can be difficult does not mean they are not worthy to attain.
1 In many categories from apparel to furniture, from Barneys, Diesel, Charlotte Russe, Z Gallerie, Payless, and ShopKo to many more. Manufacturers include Bayou Steel and Colt. — Return to article text above
2 New York Times 10/23/2019 One Family Built Forever 21, and Fueled Its Collapse — Return to article text above
3 Ibid — Return to article text above
4 How Could Travel Giant Thomas Cook Fail? — Return to article text above
5 Wall Street Gears Up For Onslaught Of Oil & Gas Bankruptcies — Return to article text above
6 In spite of talk about deregulation, demand for coal has been steadily declining due to demand from energy, steel, and so on, that has been switching to other energy sources. — Return to article text above
7 Examples: Read about Pontiac, Michigan and the troubled economic outlook in Powder River Basin, Wyoming. — Return to article text above
8 From sensors, social, weather, and so on. More on modern supply chain data. — Return to article text above
9 Crisis in demand management exacerbated by spreadsheets — Return to article text above 10 You can access this report here. And read from Complexity to Clarity on S&OP. — Return to article text above
11 Fortunately, they embarked on a path to total asset visibility and used newer systems to accomplish it. Read the U.S. Marine Corp Case Study [PDF]. (For some readers, the PDF automatically downloads without opening.) — Return to article text above
12 And companies are buying the new tech. There has been a real uptick in buying of supply chain applications of late. — Return to article text above
13 Hokey Min, Bowling Green State University, Artificial intelligence in Supply Chain Management: Theory and Applications. (You can get this report on Research Gate) — Return to article text above
14 It is always good to get first-hand opinions with lots of body language; or they may have experience labs in which to see if customers struggle to understand and use the products. — Return to article text above
15 Gamification — Launching a new product and want to know if consumers will like it? What features will they really want? What are they willing to pay for it? Engaging the consumer before a product launches or before buyers place risky bets on products can not only save companies from product obsolescence and markdowns, but can also help them set the initial ticket price. Setting the right price initially can also avoid markdowns and margin loss later on. — Return to article text above
16 A good article here breaks down sentiment methods and examples. — Return to article text above
17 University research projects funded by big business have been applying AI for over 10 years. For example, Hokey Min wrote, “For instance, Yu et al. (2002) proposed a dynamic pattern matching procedure within the agent-based system framework that combines human expertise and data mining techniques to predict the demand for new products. Their experiments indicated that the dynamic pattern matching procedure outperformed exponential smoothing techniques with respect to forecasting accuracy. In contrast to exponential smoothing, which merely relies on historical data, the dynamic pattern matching procedure utilized multiple agents to capture past (base-line agent), current (causal agent), and future (pattern agent) customer behaviors that helped improve its forecasting accuracy. Similarly, Jeong et al. (2002) improved forecasting accuracy without relying heavily on historical data by introducing a genetic algorithm-based causal forecasting technique that outperformed traditional regression analysis.”
Progressive demand planning companies have been increasingly using AI for a long time, without flashing it so much until recently, as users have become more comfortable and have begun deploying it in some rather remarkable ways. Big money is being poured into this space, which we will discuss more at length in our Demand Technology for 202X report in the new year. — Return to article text above
18 Companies like First Insight and, of course, IBM’s Watson are great examples here. — Return to article text above
19 Making Allocations Smarter. — Return to article text above
20 Read: Business Transformed:> B2B Communication Holds the Key for Success in Today’s Economy — Return to article text above
21 Developing an approach to demand signal repository to rapidly collect and analyze this data is critical. Approaches like FlowCasting have conceptually explained the benefits of getting this data, and hold promise — Return to article text above
22 We will be publishing a paper on AI in Supply Chain soon. — Return to article text above
To view other articles from this issue of the brief, click here.