The contents of those baskets are sorted by codes, fed seamlessly into powerful spreadsheets and analysed by a team of dedicated number crunchers.

By the end, Dunhumby has an eerily complete picture of individual shoppers.

It knows their spending habits, their eating habits, their hair care preferences.

It can even isolate a 20- something's contradictory tendency to buy 10 chocolate bars and low-fat milk each week, and it can hone in on the ideological outlook of an organic devotee who buys Fair Trade products.

Information

The plethora of information is fed back to Dunhumby's clients, led by flagship customer Tesco, which now has an 84pc stake in the data processor.

Then the "magic" starts, as retailers match their offerings with their customers, zoning in on the products most likely to appeal to their demographic, and using their intimate knowledge to launch special offers and promotions.

More tailored still, a retailer like Tesco can use Dunhumby's information to send personalised offers to Clubcard holders.

Those who shop for dog food every week might get a special discount on Pedigree Chum, those who buy wine religiously might get a promotion on premium cava.

The promotions, both instore and voucher-linked, are then monitored back at Dunhumby's nerve centre, so retailers can drive promotions that work and cull those that don't.

Selected suppliers are also given the Dunhumby information, allowing them to tailor their products to customers' needs, or weakness, depending on your point of view.

The end result is a "win win win", Humby says, with higher sales for Tesco, higher sales for their suppliers, and a "better shopping experience" for customers.

The process has an unnerving ring of Big Brother about it, but the mathematician emphatically denies any suggestion of intrusion.

For a start, the company only engages with customers who sign up for Clubcards in the case of Tesco, or loyalty cards in the case of Dunhumby's other corporate clients.

"If you want to have a dialogue with customers, you have to have permission," says Humby, "we have permission".

Humby also makes much of Tesco's decision never to commercially exploit Clubcard data by "selling customers" to third parties.

"The data could be commercially exploited but I wouldn't advise it," he says.

"We're very focused on the 'win win win'; we're very clear that it's about improving customer experiences. We're not selling your name and address to Procter & Gamble."

Then there's the argument that the Dunhumby-inspired insights actually benefit customers.

Humby says responses to tailored vouchers are far more positive than reactions to mass market campaigns, which ultimately means better value for customers.

"Voucher redemption (for tailored messages) is 20 to 30pc -- mind-numbingly high," he says. "If something relevant is sent to customers, they're much more likely to use it."

Tesco's "Supplier Insight" programme is largely used as a bargaining tool, with the supermarket giant offering "blue data" to suppliers tracing general trends with no personal information, but Tesco also uses the scheme to encourage promotions for frequently overlooked customers like older age groups, in a nod to social responsibility.

Humby also stresses the "fantastic security" surrounding Dunhumby's web of data, ensuring one's Jaffa Cake or Coco Pops habit will never be exposed by a stolen laptop.

Customers aside, the Clubcard project is widely recognised as one of the driving forces behind Tesco's rapidly expanding empire, which racked up £2.8bn in sales during its last financial year.

Loyalty

Humby said Tesco was "about three years ahead of the curve" when it began seriously data mining in 1995, aided by the skills of Humby and his wife Edwina Dunn.

"Tesco without Clubcard is unimaginable," he says. "The loyalty data is such a valuable set of insights that it's used with a light touch everywhere."

Certain gems can be gleaned instantly.

Tesco has long known, for example, that promotions on coffee don't work because people tend to be loyal to one brand, while beer promotions are a winner because shoppers are happy to try many varieties.

Using such insights, displays can be changed within days, ordering can be changed in about two months, and entirely new products can be commissioned in a little longer.

"It doesn't replace the buyer or the marketeer, it's an empowering tool," Humby adds.

In recent years, other retailers have gotten in on the data mining action, most notably Tesco's key competitor in the UK, Sainsbury's.

Yet Humby stresses that while data mining works well in "high transaction businesses", it will never hold the key to retailing success.

"If you're stores aren't right, if you haven't got the prices right, then people won't shop," he says, "You can't exploit a loyalty card without the fundamentals."