At M&M Food Market, the most valuable product isn’t necessarily the food, but the mix of different ingredients the company uses to grow in Canada’s highly competitive grocery and prepared meals sector.
The recipe isn’t complicated: Take a well-known brand – M&M Meats until the company rebranded in 2016 – add a helping of customer data and sprinkle it with predictive analytics and artificial intelligence tools.
The result? A company that has gone from a niche purveyor of bacon-wrapped filet to a data-driven prepared-food business tuned into customer habits and desires.
M&M’s mission today is to be agile and attentive to the appetites of millions of regular customers across Canada, says its chief executive officer, Andy O’Brien.
“People’s eating habits have changed, and we had to become more relevant,” Mr. O’Brien says.
Founded in 1980 by Michael Voisin and Mark Nowak (“M&M”), the company’s makeover picked up steam after the company was acquired in 2014 by Toronto private equity firm Searchlight Capital Partners LP and when Mr. O’Brien’s new team came on board.
The new management team quickly discovered that M&M was sitting on a mountain of data – or about 18 years’ worth of information on roughly nine million customers, both regular and casual purchasers of M&M products.
“We had about 40 per cent of Canadian households in our database, but we weren’t using this information,” Mr. O’Brien says. “This is information that could build value for the customers and also build value for us.”
The company began to drill down, including looking at how peoples’ eating patterns have changed.
“The average commuting time [in Canada] has doubled since the 1980s, so they’re getting home later,” Mr. O’Brien says. “People in households eat differently too – they have separate meals at different times.
The company also noticed people were preparing meals differently. “They’ll take salad ingredients, for example, sprinkle some cheese on top and mix this with prepared chicken. They’re more creative, but they look for ingredients in a ‘steady-state’ that they can rely on, and that makes frozen ingredients more relevant,” he says.
To make better use of this data, M&M hired the California technology firm Retention Science (ReSci), which applies AI to customer data to analyze the information consumers offer about their likes and dislikes.
“We were careful to shield people’s identities, while we downloaded to them something like 25 million transactions,” Mr. O’Brien says.
Together, M&M and ReSci used AI to build a predictive model that could indicate what consumers liked best, what they would likely want to buy next and what kinds of food products that are likely to become trends – for example, prepared gluten-free foods.
The AI model offers a nuanced picture of consumer behaviour.
“For example, we know that people in certain areas buy our chicken more often than in other areas – we can let them know when new chicken products are available or when there are sales,” Mr. O’Brien says. “Other people might buy our pot roasts and then not buy them again for a while. Yet they tell us they like it, so we can remind them. It works.”
Predictive science can help companies determine how much their customers are likely to spend in the future on their products, as well as “whether and when there’s a risk they’ll be leaving you [and] what products they recommend,” says Derek Kwan, ReSci’s chief operating officer.
It’s a two-way street, Mr. Kwan explains. “AI and predictive science helps customers find the products they like best, and the company can tailor its products and marketing to what the customers like.”
The results are noticeable, Mr. O’Brien says. Since engaging ReSci, the average M&M customer’s food basket has increased by more than a dollar, profit margins have risen and sales per customer have gone up.
Customer engagement – consumers talking to or e-mailing the company – has also increased by 15 per cent, Mr. O’Brien says.
Deploying AI has also helped M&M organize its marketing more strategically because it understands its customers better.
“We’re opening new stores, but we also see areas that are too small to support a store, but where we notice customers who would like to buy our products. In those areas, we partner with local merchants and install branded shelf space in their stores,” Mr. O’Brien says.
Perhaps most importantly, the decision to deploy AI enables a mid-sized company like M&M, with 95 employees, to compete with grocery giants such as Loblaw Companies Ltd. or Sobey’s Inc., says Stephen Thomas, director of the programs in management analytics and AI at Queen’s University’s Smith School of Business in Kingston.
“Using AI can make or break a business these days,” Mr. Thomas says. “It’s important to small and medium businesses; I see local businesses that don’t have a system to gather data at point-of-sale, and I worry for them.”
“Having good data and using it well lets companies and managers tailor their businesses. Everything they want to know can be at their fingertips,” he says.
If anyone is worried about “singularity,” when robots take over all human functions, including thinking and managing, Mr. Thomas says they shouldn’t. AI might analyze and predict what consumers might want to buy for dinner, but it’s not good at anticipating the emotional snap judgments and whims that cause consumer trends in the internet age, he says.
“We still need good managers with ideas and judgment for that. People still matter,” Mr. Thomas says.