Since its inception, Loblaw‘s PC Plus loyalty program has been a data play.
The grocer uses data points from every single purchase to fuel the hyper-personalized offers it sends out every week via PC Plus, effectively creating thousands upon thousands of individualized flyers.
As the loyalty program’s devotees have come to expect, PC Plus offers aren’t based on what’s overstocked or being pushed by packaged goods brands; they’re chosen by an algorithm that learns what each customer buys, then caters to their personal tastes.
Speaking at the Canadian Marketing Association’s recent CMAinsights conference, Loblaw senior director of loyalty and customer analytics Peter Danforth explained the strength of PC Plus comes from its strong technological base.
Early on, Loblaw spent considerably on developing the technology that runs PC Plus, dodging what Danforth said is a common problem for companies – executives get excited about campaigns and put the message before the technology. With PC Plus, Danforth said Loblaw tested, adapted and, perhaps most importantly, invested in getting the offer algorithms right before taking the program to market.
Early on in development, store data made one thing very clear: no two customers are the same. Loblaw tried grouping customers, but found the variances in shopping patterns were too vast to deliver meaningful offers on anything other than a one-to-one basis.
This marked a major change in how the grocer delivered its promotional offers. In the past, stores could make decisions at a local level about what to promote based on supply and demand. That still happens through other avenues such as flyers, but with PC Plus, Loblaw decided to trust the data and shape offers based on it showed.
Danforth said this process has been key to success for PC Plus. He offered an example: at Loblaw’s upscale Maple Leaf Gardens location in Toronto, the company traditionally wouldn’t think to make offers on a product like No Name canned vegetables. But if the purchase history suggested a customer was likely to buy No Name canned vegetables if presented with an offer on that product, why would the grocer offer something else?
“You have to listen to the data,” Danforth said. “You can’t tell the customer what they want.”
The only way to truly do personalized marketing, Danforth said, is to curate deals and offers from an entire product line. If beef sales are doing well at a store, for example, there’s a temptation to offer deals on chicken. That’s the wrong method, Danforth said. If the data shows a customer prefers and is more likely to purchase beef, an offer on chicken may be a wasted opportunity.
It took time within Loblaw for the company to adjust to this mindset. Danforth admitted there were people inside the company who struggled to work with its data scientists, but he said getting executives and the marketing team to collaborate with the data team was integral to making PC Plus work.
For marketers especially, he said it’s crucial to collaborate with data scientists and understand the work they do. “If you don’t have a direct line to the data scientists who are doing your personalization for you, they are running your marketing program – you just don’t know it,” he said.
