Written by 5:02 pm Big data supply chains, Change Management, customer-centric supply chains, Demand, Demand driven

Dealing With The Supply Chain Gloppy Mess

During August, I moved and I spent the weekend painting my new shed with gloppy, wet deck paint. As I methodically pivoted the roller to apply the stain, I made quite a mess. Standing in the hot sun covered in paint was a good time for reflection. (Deck stain is not designed to be rolled onto a ceiling.)

This past weekend was a US Holiday–Labor Day. For the first in seven years, I was not heads-down preparing for The Supply Chain Insights Global Summit.  COVID-19 drove a cancellation, but I hope to host the event again in 2021 in Franklin, TN. Covered in the paint spatter on a beautiful Sunday afternoon, I thought about what I would have said to the audience if the event wasn’t canceled. Here I share these insights.

Dealing With A Gloppy Mess

During this pandemic, sensing market changes, using data, and driving decisions at the speed of business matters more than ever. Companies are drowning in data and short on insights. Traditional processes aren’t equal to the wild swings of the market. Pre-pandemic only 30% of supply chain leaders were satisfied with their supply chains, and during the pandemic, business leader satisfaction is falling precipitously. Most are struggling to find consultants to help. Let’s look at some of the market trends to understand the severity:

  • Lumber prices skyrocketed 60% in the last six months. The impact? A moderately-priced new home in the United States now costs an additional $16,500. Builders in my area are pre-paying for lumber to hold prices at the cost of $250,000/house. When will lumber prices return to normal prices? No one knows. Tarif issues on Canadian spruce lumber complicate the surge in demand for the building surge. Who would have thought that new home building would surge in the face of record unemployment?
  • Food and beverage companies are experiencing double-digit demand. Most are capacity constrained. For example, Campbells Soup is struggling to keep up with the demand for soup, potato chips, and cookies. Sales are up 12%. Supply is the challenge. For years Campbells strove to grow the soup category only to lose share in the soup category due to limited manufacturing capacity for the Swanson’s brand.
  • Demand for electricity and natural gas is down 20-25% globally as remote workers shelter in place and operate offices in their homes. Renewable energy usage is surging and oil is fluctuating.
  • Retail dock doors were a formidable constraint as grocery sales increased 30%. Scheduling the doors and receiving products at the speed of consumption slowed movement to the shelf for everything from eggs to toilet paper. What is needed is a redesign to move more products to the market.
  • The pending COVID vaccines need a cold chain to move the product to the market. There are not enough containers in the world to meet the pending demand. With air capacity limited due to the decline in air travel, companies moving products using cold chain assets will struggle.

This list of market shifts is endless. Every day, there is a new factoid supporting that the market is teaching us a lesson: the traditional focus on marketing-driven and sales-driven processes is no substitute for market sensing and demand translation. The focus on IBP in Sales and Operations Planning (S&OP) and Demand-driven MRP (DDMRP) make us less market-driven. (DDMRP is based on the use of math to determine material buffer strategies based on order patterns. While useful when the order patterns are stable, it is less helpful in times of gyrations.) Similarly, demand planning based on order history is not equal to the challenge.

Over the course of three decades, companies invested in enterprise systems to improve functional efficiency. Today, the pandemic is highlighting the folly: efficient functional systems tethered to ERP are inflexible and self-serving. The conventional supply chain processes are unable to sense and respond as markets shift. The market is shifting with wild gyrations and companies are not equal to either seize the opportunity or mitigate the risk. The attempts to shorten cycles increases noise.

In short, the order is not a good proxy for market demand. As a result, during the pandemic, all processes based on order pattern sensing are obsolete. The reasons are many:

  • Short shipments and out-of-stocks. Most companies are not honest with themselves on current performance and the growth potential to fill orders and sense/respond to true consumption.
  • Rising Complexity. Product and channel complexity increased the tail of the supply chain increasing demand latency. (Demand latency is the time from channel purchase to order receipt.) Order latency for a fast-moving product in a large channel averages twenty days while a slow-moving product in a lower-volume outlet is over one-hundred days.
  • Changing markets. Lockdowns and sickness swing markets in the pandemic. Consider the impact on elective surgery and protective supplies. Sensing at the speed of business is more important than ever.
  • The rising cost of supply. The dramatic shifts in the market result in supply outages with price volatility. A good demand signal is essential to forge competitive buying strategies.
  • Constrained logistics. Pre-pandemic if the demand signal was poor, air expediting was expensive but an option. Today, the bellies of the airplanes are full, and the cost of air freight is too high for many manufacturers. In the fall, reefer container capacity will be non-existent with the expected release of the fall COVID vaccines. We have never operated in an environment with the constrained logistics looming on the horizon.

In today’s IT architectures there is no place to put and use market signals–Internet of Things logistics signals, weather data, consumption data, rating and review insights, traffic patterns, and unstructured data. We have tethered the supply chain to transactional enterprise data. In short, today’s supply chain world is a gloppy mess. While many companies attempt to improve the response by being reactive on traditional data sources, this is not the answer.

What To Do?

Let me start with what not to do. I had a call a month ago from a supply chain leader that I admire asking for a shortlist of consultants to use to formulate a project for the next-generation supply chain. Let’s call her Mary. In the call, I challenged her logic:

  • What is a next-generation supply chain? Why does it matter?
  • Why do you believe that you are failing with your current technologies? Recent investments are not successful, can you tell me why?
  • Can you share why an RFP approach makes sense?

We agreed to agree to disagree that an RFP-based project approach could drive success. Since I like this business leader, I agreed to build a short-list of inciteful leaders. I mulled-over this list for two days. I struggled to develop a listing of consultants that I could recommend. (I find most consultant briefings are full of high-level hyperbole, acronyms, and high-flying promises. The briefings are full of terms like demand sensing, control tower concepts, digital transformation, and cognitive computing, but when I ask for definitions, the answers are not forthcoming. Mind-numbing…)

In short, now is not the time for a RFP. In contrast, one of Mary’s competitors is piloting a cognitive computing platform to sense baseline lift, redefining assortment by region, and generating monthly demand shaping plans by individual retailer. The project has no consultants and the ROI is in days.

One of Mary’s other competitors is implementing SAP HANA and a packaged order-based forecast technology and is struggling to read the market. This company deployed a team from Palantir to discern patterns of demand insights, and this work sits neatly in binders on the VP of Supply Chain Management’s desk. The problem? It is not sticky. There is no place to put the output of the work into the conventional enterprise architectures. As the market changes, the consultants are slow and expensive to understand the market drivers.

A Shift From Planning To Decision Support

Underneath the technology market in advanced analytics is the move from planning to decision support. The focus is on building outside-in processes. Let’s start with some definitions:

  • Planning. The use of advanced math and optimization to improve outcomes. Planning is very “engine centric.” The mathematical engines are fed data, there is a step for engine processing and then the output is written to a system of record for use by business users. The planning engines serve a well-defined group of business users, but there are not connected to an adaptive strategic model or a set of cross-functional processes. This list can go on-and-on, but includes Revenue Management, Sales Account Planning, Market-Mix Modeling, Price Management, Category Management, Demand Planning, Budget Forecasting, Inventory Management, Tactical Supply Planning, Transportation Planning, and Commodity Management. In the evolution of decision support, these applications become legacy. However, this will not happen quickly.
  • Decision Support. The decision support technologies combine advanced math, flow analysis, policy execution with adaptive rule-sets. The evolution spawns a group of technologies that are outside-in and cross-functional. The focus is on improving decisions through bi-directional orchestration market-to-market using the channel insights, logistics signals, and sourcing market data.

In the transition shown in Figure 1, companies will struggle. The first step will be to hire data scientists and challenge them to use techniques like Python and R to improve today’s functional outcomes. However, the management of data scientists is not the fix that is needed. The gap? There are many, but the largest is the need to move from inside-out to outside-in processes. We must get past the functional data/process silos to serve the market. The second step will be to hire third-party experts like AimPoint Digital, Mu Sigma, Palantir, or Tiger Analytics to sense and translate market data. These groups will be successful in driving insights, but not at the speed required or a sustainable cost/value relationship. The third step is to overlay machine learning or advanced analytics over existing data models. This issue is that the data models are enterprise-centric unable to use market data. As companies move through these steps they will start to knock on the doors of technology innovators.

Figure 1. Evolution of Decision Support

Lights-Out Planning Is A Bridge Too Far

As this work continues, business leaders will lose traction chasing shiny objects. Companies will seek solutions for autonomous planning or applications that can automate the sense and respond capabilities. My caution? Take slow steps to understand how to define process capabilities. If not, you will automate historical processes that are not adequate for the challenge.

What to do? Work hand-in-hand with technology innovators (my preference is to not layer the project with consultants and define the art of the possible. Start by listing the gaps, or black-holes of the supply chain, and explore all forms of data within the organization that offer possibilities. Make team members define each term. (Terms like visibility, control tower, demand insights, forecast, demand sensing, and planning have many different and varied definitions.)

In parallel, define the rudder, or compass for the project. his is a set of measurements to gauge improvement: out-of-stock improvements, forecast value added, schedule adherence, total cost, and bias/error.  (This list can go on and on, but tether the groups’ work in business deliverables that are not functional or political.)

I miss being able to lead this discussion at the Supply Chain Insights Global Summit, but I look forward to helping you to get past the gloppy mess to drive value. I welcome your thoughts.

Close