Opinion: AI mixes it up in the grocery aisle

I  must give writing props to Missouri State University’s Office of Strategic Communications. It’s not every day, or week even, that the first sentence of a news release hooks me. 

News release headlines and email subject lines, yes. I know in a matter of seconds whether I’ll read on, move it to a coverage folder, share it with my journalist colleagues or send it to the trash bin. But I regularly get emails from this MSU office, and while interesting, they’re mostly feature-type content that’s not an ideal fit for Springfield Business Journal. 

I received one from MSU last month that caught my attention in a different way. The first paragraph read: “Cake mix, frosting and artificial intelligence. These ingredients go together smoothly for Missouri State University associate professor of information technology and cybersecurity (ITC), Dr. Lawrence Yang.”

There are many university professors doing meaningful research but not all have a business impact. This one does. How?

As the release explains, Yang’s research can assist business decision making in an innovative way.

Let’s go to that grocery aisle, where we’re looking for that cake mix and frosting. We want to bake a cake, obviously. But what does that grocery store owner need to know about our consumer wants – and more importantly our behaviors – once we’re in the store? Here’s the short answer from the release: AI can predict the items that land in our shopping cart.

“I am interested in applying artificial intelligence to help solve business problems and better predict possible outcomes,” Yang says in the release.

Yang is really analyzing the relationship between two products that are usually purchased together. But it’s AI doing the legwork. It’s work that has been done before in the realm of cross purchases – think about buyers of detergent and dryer sheets or a hairbrush and blow dryer. These relational buying habits, particularly considering price fluctuations in one product compared with its counterpart, have been studied in what’s called econometrics. For instance, does a price increase or decrease in one of the items affect the sales volume of the other?

Yang points out a flaw in the traditional field of study. It has to do with scope and speed.

Econometric models can only handle so many variables before the data break down and cause delays in the outcomes. Not so with AI.

According to MSU, Yang has spent six years and authored or co-authored six publications of research centered on how predictive technology can make businesses more efficient in decision making.

Fascinating.

Here’s a summary of his case study (from highlights published in an MSU magazine called Mind’s Eye):

Yang’s research began with a raw data set of 20,000 purchases made publicly available through the Nielsen Consumer Panel. Then, comparison tests between the traditional models and AI started to reveal his conclusion: Yes, the test performance and results were similar. But the time to achieve them was drastically different: two weeks versus two hours. And when just a handful of variables was added to the econometrics model, he saw signs of it becoming overwhelmed.

“I believe AI has the ability to affect organizations’ decision-making skills in a positive and efficient way,” Yang told the campus magazine.

Here’s one application in the grocery store example that I could see: If the price of one of the complementary items increases on the supplier side, the seller might consider eating a portion of that extra cost in order to not lose out on sales of the other item. It’d be better to lose pennies or dollars on one item than to lose the sale of both items altogether. Having certain information quicker could produce real-time decisions that save profits. And if the theory is not immediately paying dividends, then the store owner could make the real-time decision to pass along the price increase and see what happens. But, to Yang’s point, they’d have predictive data to guide those decisions.

It’s a new world, with tech at the forefront, and AI is increasingly expanding into many areas of business. Who knew AI would be behind the scenes determining the best prices for the most sales at the local grocery store?

Professor Yang, that’s who.

Springfield Business Journal Editorial Vice President Eric Olson can be reached at eolson@sbj.net