Across 388 New Zealand supermarkets, ShopIt has observed that shoppers are more willing to switch between retailers than brands when AI identifies better value.
Once shoppers can compare across chains, they do. Most ShopIt subscribers pick stores from more than one chain straight away, and many pick all three. In most of the comparisons, the cheapest option turns out to be somewhere other than the shopper's usual store.
“People act on that. We see baskets sent through to all three supermarkets' online stores, and a fair share of the time the app suggests splitting the shop across two stores rather than doing it all in one place,” said Jason Quigley, founder of ShopIt and feedi.
“People do weigh up the drive, though. One user told us where she lives, that she doesn't pay for petrol, and asked where the cheapest place for nappies was. That's the calculation people are making.”
At the same time, brand loyalty exists only for a very short list. Quigley highlighted that nappies, laundry powder, soft drink, cereal and bread are asked for by name, whereas almost everything that goes into a meal isn't.
“People ask for mince, pasta, Greek yoghurt, and they take whichever is cheapest. They'll even volunteer swaps, like "use pork mince instead of beef if it's cheaper." So shoppers aren't loyal to brands; they're loyal to a handful,” he said.
For example, one shopper in East Auckland showed most of this in a single visit. She kept her shopping list in a spreadsheet, fifty rows with quantities, and rather than type it out, she photographed the screen and sent it with "this is my shopping list."
About a third of the rows named a brand: Weet-Bix, Marmite, Sealord tuna, Iams and Fancy Feast for the cat, Angel Bay patties, Oreos, Nuttelex. The other two-thirds were just "chicken breast 6kg", "cheese 1kg", "eggs 60", "olive oil", "chilli beans", "burger buns".
She'd also written her own rules into the list: "beef sausages, 6 pack x3 if on 3 for special, otherwise x1", "Angel Bay patties, preferably lite", "Greek protein yoghurt, either Anchor or Yoplait".
The app matched 44 of the 50, told her which six it couldn't find, and recommended splitting the shop: chicken, cheese, eggs and produce at Pak'nSave Sylvia Park, and the club-card items like the tuna, cat food and mince at Woolworths Pakuranga, roughly NZD 28 cheaper than doing it all at one store.
For the two-thirds she hadn't named, the AI picked the brand, and she didn't change a single one. She said "split between both stores", added a few things the list had missed. She has done the same thing several times since, and the pattern holds each time: the brands she named stay the same, and the rest move with the price. Her cheese went from Rolling Meadow one month to Pam's Value the next because it was a dollar cheaper, and her tuna moved stores because Pak'nSave had it at NZD 1.39 instead of NZD 2.20.
Quigley also mentioned that in ShopIt’s case, a brand only shows up if it wins on price, fits something the shopper has said they need, or the shopper asks for it by name, and that product positioning and packaging don’t factor in, so for most of the basket, the choice has already been made by the time the shopper gets to the supermarket's website.
On how the technology decides which pasta, sauce, meat, or cheese to recommend based on price, promotion, previous purchases, dietary requirements, availability, or brand preference, Quigley said it depends on how it was built.
Read more about AI & the trolley in the latest issue here
