Skip to main content

Select your location

Supply Chain Management: Taking the reins with data

Mark Ardito
shipping containers in yard

It was fun while it lasted, but it didn’t prepare us properly. Since the introduction of just-in-time logistics by Toyota in the 1970s, manufacturers in nearly every industry have grown accustomed to the near-perfect operation of supply chains with a seemingly infinite supply of raw goods, fast transport, and apparent efficiencies.

Even large companies had no problems managing it all by way of Excel spreadsheets. Meantime, on the consumer side, top e-commerce players like Amazon Prime have led consumers to expect 1-, 2-, or even same-day delivery of their orders.

Everything has changed, due in large part to the COVID pandemic. Our supply chains, which we thought so reliable and resilient, were revealed as the fragile and complex networks that they really are. As COOs and procurement leaders continue to tackle problems as they seek ways to fill their supply chain with goods, CTOs and data scientists are stepping in to help.

I was recently joined by Bryan Frank, Director, Strategy Practice Lead at Kin + Carta, to talk about the present and future of supply chain management with a roundtable to tech leaders. It was a lively discussion with lots of questions, insights, and ideas shared between participants.

Here are some of the key points that emerged:

Move from passive to predictive data 

Dashboards are a steppingstone but not the destination. Once seen as the goal of supply chain management, they are in fact not so helpful in today’s environment. The main reason is that dashboards deliver information passively, too late for effective remediation.

As Bryan pointed out, “When your dashboard tells you that chicken didn't arrive at the food distribution company, and then McDonald's, Burger King, and Chick-fil-A never got chicken delivered to them, it's already too late.” Predictive analytics is key to truly effective supply chain management.

By leveraging AI solutions, such as Octain, organizations can build predictive, explainable models that will accelerate data-driven decision making and digital transformation outcomes. This enables companies to make better predictions for future demand and get ahead of supply chain shortages and blockages.

Predictive data means operating in the cloud

Effective predictive analytics requires large amounts of data processed continuously and quickly. And in AI and machine learning, It’s clear that this can only be done in the cloud. The amount and space requirements of on-prem assets required to use predictive data are prohibitive.

Research shows that 54% of companies are increasing or substantially increasing their investment in cloud computing and storage to make the supply chain more resilient.

Business lines must be on the journey too

People are always a key ingredient in any technology initiative, and the more significant the initiative, the more buy-in is needed. Discussion participants were mixed in terms of the support they receive from business leaders for technology-enabled supply chain management.

When the business is sluggish in understanding the magnitude of the impact of technology-enabled supply chain management, tech leaders must pursue strategies to gain interest, support, and willingness to collaborate in making necessary changes.

Data science is not a single skill set 

Data science is fast becoming an integral part of many companies, but the term is widely used in technical discussions as a single skill set or person. In fact, data science is an umbrella for a range of skills and activities.

Rarely does one data scientist have all the experience needed to deliver the best results. It usually takes a team of people, each member focusing on one aspect. Keeping this in mind when ramping up a data science function will support success.

Technology-enabled supply chain management

Whether you’re a supply chain giant or a consumer brand upstart, there are certain steps you need to take to start minimizing risk and maximizing market potential as part of the global supply chain. 

A good way to begin standing up technology-enabled supply chain management is understanding where to start:

Concerned with speed and agility in the supply chain?

Learn more

Start with a small effort as a test case and go from there

Just start somewhere! Use a small dataset or a single location as a first test case. Verify your plan, adjust, use it as a business case if needed. From there you can scale as it makes sense.


Buy, don't build

This question accompanies almost all technology initiatives. And with almost all technology initiatives, the answer is “it depends.” What part of the supply chain management picture are you talking about? I talked a bit about our work with clients on route optimization, for example, which is an area where good tools already exist. Generally speaking, if buying is a good option, take it.

Consider consolidating

This suggestion came out of a discussion about warehouses and a company that is actively acquiring other businesses. Consolidating and possible reconfiguring the warehouse map, for example into a hub-and-spoke set up, can help supply chain management in general and technology enablement in particular.

Run experiments

This is especially applicable to incorporating AI and machine learning into your strategy. Rather than trying to develop some set of heuristics or business rules based on the experiments, train some models to make those decisions for you over time. And it is completely okay to fail.

We find that we are constantly coaching our clients that not only is it okay, but failure is expected as part of the experimentation process. That’s how new learning occurs and improvements get made.

In an environment where Excel is still used for supply chain management, even by large companies, technology leaders are stepping up to make changes. Through a combination of creativity, data management, the cloud, and AI, supply chains can bounce back and even end up significantly improved.

Share this article

Show me all