How Machine Learning and Data Lakes Help Your Pets (and Your Business Too)

David Colyott, Supply Chain, Mars Petcare

David Colyott, Supply Chain, Mars Petcare

These days, most of us encounter machine learning and the power of big data regularly. Whether its Alexa learning your music preferences or chatbots offering customer service, technology affects transactions in ways we barely notice any more.

But are we doing enough with it in the supply chain industry? That’s a question we ask ourselves regularly at Mars Petcare.

The answers can sometimes be abstract, focused on things like capabilities and elasticity. But when you stop to pick up cat food, if Fluffy’s favorite brand is not on the shelf, there is no abstract about it—your furry best friend is going to be disappointed.

To make sure that does not happen, we are applying new technologies to make our operations even smarter. Here are two examples of advances we are currently pursuing.

Using Machine Learning to Automate

Machine learning is not a new concept, but it offers new ways to automate tasks and simplify processes. At Mars Petcare, we apply it to our transportation management system to make administrative tasks faster and more effective.

For example, a lot of our customers manage schedules for warehouse deliveries through a customer portal. Securing optimal appointments mean the best opportunity for on-time delivery, so we use machine learning to automatically secure appointments at the earliest opportunity within the communication process.

This frees up a significant amount of time for our team. Depending on a customer’s tools, in many cases our system can learn to watch for a trigger email and go immediately to that customer’s portal to schedule a time. There is no user intervention required.

Similarly, in our Customer Care group, like any company, we do a great deal of order management. Our system has traditionally been high-touch, with customer service associates opening orders and updating them, such as adding weight, substituting an item, etc.

Now, we use machine learning to automate the highly-repeatable activities, making them lower-touch. Our team can focus more of their time on added-value, customer-centric activity, and that means more satisfied customers and, hopefully, happier pets.

Securing optimal appointments mean the best opportunity for on-time delivery, so we use machine learning to automatically secure appointments at the earliest opportunity within the communication process

Why it matters: A lot of times when people hear about machine learning, they worry it will take away human jobs. But we see the opposite opportunity. We are a growing company in a growing category. Machine learning helps us maintain that growth without adding exponential resources. It is all about enabling growth.

Enabling cost to serve insights with data

An even bigger opportunity is bringing the power of big data to understand the cost to serve. Most large CPG companies face challenges in managing their portfolio. They have many brands, many different origin points across the supply chain, and a lot of SKUs within them. They might have legacy systems managing data in different ways, and in different places.

We face similar challenges at Mars Petcare, where we have nearly 20 points of origin for shipping across the U.S., both company-owned and third-party distribution, and numerous leading brands like PEDIGREE^®, IAMS^™, GREENIES^™ and TEMPTATIONS^™.

This means a lot of disparate inputs. The data from our transportation management system, warehouse management system and SAP platform is combined into a shared data lake. Our analytics team can then create tools to understand the cost to serve, and support decision-making that takes into account all elements of the supply chain.

For sales, this means opportunities to create joint business plans with customers based on actual data about the products they want, and the relevant supply chain enablers or constraints. For marketing, it means supply chain inputs that can aid in annual marketing plans, for example, offering insights to help determine the optimal time for a new product launch.

And for pets, it means more access to the foods they love.

Why it matters: It is one thing to know how one location in your supply chain is performing. It is a lot more powerful to get all your data into one place and be able to use insights across your whole network. If you are not certain what your biggest supply chain advantages or constraints are, cost to serve is a key to unlocking those insights.

What You Can Do to Get Started

In many organizations, conversations about data and insights happen only at the department level. But then you miss the bigger picture. When you bring together Sales, Marketing, Demand Planning, and Supply Chain, it enables total enterprise insights and solutions, and pulls all the pieces together.

We work hard to do this at Mars Petcare, and one of our strengths is our collaborative culture and commitment to uncovering insights. We have launched a new division, called Kinship that is dedicated to realizing the promise of insights for pet health and well-being.

Whether it is a supply chain transaction or an analysis of pet activity, at Mars Petcare, we see the potential for data and technology to transform, and help us deliver on our purpose: A BETTER WORLD FOR PETS^™. That is good for our supply chain and our industry. And yes, it is good for Fido and Fluffy too.