Interview by Symphony RetailAI’s Paul Hoffman. At the annual Xcelerate Retail Forum 2019 conference Jason Burnett and Jeanette Buck shared lessons learnt with AI and how together with changing shopper demand across the companies’ four channels is enabling SpartanNash to think differently about its replenishment strategy and inventory deployment.
SpartanNash is an American food distributor and grocery store retailer headquartered in Byron Center, Michigan. The company’s core businesses include distributing food to independent grocers, military commissaries, and corporate-owned retail stores in 44 states, Europe, Latin America, and the Middle East.
What do you think is one of the biggest challenges in inventory management?
Keeping up with our customers’ volume trends. Over the past few years, we have been seeing volume shift from our traditional wholesale business to our national accounts, which is directly related to consumer trends. A few of these very large national account customers are experiencing exponential volume growth. We’re staying in front of it using Symphony channel functionality, which enables us to segment that business, provide feedback to those customers and better align future expectations.
Why do you think the volume has increased so much?
The larger customers are in non-traditional formats. They are putting smaller stores in rural communities, operating online and offering home delivery. This is where the majority of the traditional volume is going. Everybody’s time is valuable, people want the convenience of click and collect or home delivery services.
Give me a sense of the complexity in terms of demand forecasting for your different customer bases.
Because of the changing volumes and varying needs of our customer bases, from a supply chain and infrastructure perspective, we have to think differently about how we’re executing our replenishment strategy and inventory deployment. In a traditional wholesale environment, things are thought of on a weekly basis and weeks of stock. With things happening with interest rates and inventory, tiering costs etc., we’re really starting to think more down into the days of cover versus weeks of cover. By having an accurate daily forecast, we can provide more of a just-in-time inventory into our facilities, which allows us to push more volume through the same size distribution network. Our strategy is more of a global view from a network perspective.
Is there a fundamental “truth” or something you’ve come to realize in your work over the years?
If you come across something that doesn’t look right, once you dig into it you can figure out what’s going on. I think the platform that we are migrating to with artificial intelligence demand forecasting and AI warehouse replenishment is going to be a true game-changer. We’re on traditional demand forecast now, but the efficiencies that the AI brings allows us to crunch more numbers and become more accurate, which plays into the whole network strategy.
The other thing I would add is that simply having a quality forecast dashboard – one that calls out the items that need to be addressed – that’s a major advantage. The system we have today doesn’t do that. Everything we’re moving toward is much more reactive due to AI technology, but more importantly, the work flows are exception driven versus having to work the whole pile. That allows you to be much more efficient; it’s the scalpel technique versus the shotgun technique. The systems tell you what needs to be addressed. Major time savings.
Please give a summary of what you both talked about at Xcelerate 2019.
When I presented at Xcelerate back in 2017, I talked about how we had an extensive amount of data. We had all these customer transactions, but we lacked the utility that had the right exceptions management and process to distill the data into usable information. Now, we’re at the next step, using our data to the fullest and gaining insights that will have major impact on business, putting our data to practical application. The channel forecast will be put into application from a procurement standpoint, a sourcing standpoint, collaboration with our manufacturers and supply chain partners. The forecast will be used for labor-planning, network planning, and how we can drive more volume through the same network.
Access to this data allows us to provide feedback to our diverse customers because we’ve channeled our demand. This enables us to be able in provide insights to the trends we are seeing about their assortment, and trends within the other channels that may enable them to take advantage of those trends.
Finally, we talked about the results that we’ve seen from the pilots that we’ve done with your AI products, and we’ll cover our continued journey as we move to migrate everything into the AI products, which would be AI demand forecasting and AI-driven warehouse replenishment.
Please tell our readers something that you like to do outside of work.
Jeannette: I really try to live it up and enjoy all that I do outside of work. I love to travel and have done a good amount of it. A few of my favorite trips have been to Egypt, Turkey, and the Greek isles. I enjoy exciting activities – I’m a bit of a risk-taker…I’ve skydived, for example.
Jason: I love traveling and fishing with my kids – I’m a big fly fisherman. Something that may be a little unique is that in 2017, we bought a lake house and we are remodeling it entirely ourselves on the weekends. Three weekends ago I put in a 60-foot beach at the house, which was 20 tons of sand.
Find out how AI-enabled Demand Forecasting is helping retailers improve accuracy and reduce manual guesswork.