The co-founder and chief executive of Gail’s has revealed the bakery chain uses a computer algorithm to choose locations for new branches.
Tom Molnar explained that his technology team has created a model that “learns on its own” after various inputs of different degrees of importance have been fed in, such as on the competition in the area.
The information gathered from “using the science” via what he described as the “same kind of thinking” as AI then helps the upmarket chain pick new neighbourhoods for their next cafés.
Mr Molnar’s revelation comes as branches in a seaside conservation area in West Sussex and Walthamstow, east London, drew backlash from locals.
Speaking on the Hungry podcast on its Monday episode, Mr Molnar said staff have been building the system over eight years, entering data such as the distance from transport links and the number of churches and schools in the area.
The key question to be answered is, “Why would somebody want a bakery?” he said, adding: “We end up with an algorithm that can tell you kind of every postcode in the country, what we might be able to do there, based on what we’ve done.”
Last year, in the seaside town of Worthing, the arrival of another Gail’s sparked a debate, with some welcoming the bakery as a sign of much-needed progress, while others saw it as a harbinger of gentrification, poised to strip away the town’s local businesses.
Meanwhile, in Walthamstow, residents were up in arms in the summer over the impending arrival of a new branch on their high street, with a petition against the opening gaining hundreds of signatures. “Preserving Walthamstow's high street is not just about supporting local businesses – it's about protecting the unique identity of our community,” it read.
Mr Molnar explained to Dan Pope, host of the Hungry podcast: “We want to go to a place that is competitive. So you look at competitive landscape, you look at the footfall, you look at the access to transport, you look at the spend that these people have available to them, so demographics. And we've put them into a model that kind of learns on its own. So every new site I have, I put in there.”
He added: “AI is more advanced, but it's the same kind of thinking where you just keep on feeding a model and the model can tell you ‘you might look at this area, I think that this might work’. And then that gets us that beginning and we can say ‘okay, let's use the science, now we need to use our brain’.”
