Friday, November 12, 2010

Beer & Diapers: Tesco Defensive Marketing

What's In Your Shopping Cart?


Businesses want to know what customers buy and potential order add-on opportunities. Supermarkets are particularly interested in their customers' behavior because of their high inventory turnover, spoilage and low profit margins. While managers will intuitively market hot dog buns with hot dogs, burgers with charcoal, or peanut butter with jelly, many businesses are turning to more sophisticated analytics hoping to differente themselves and their profit margins.

Beer and Diapers


A WSJ article from 2006 states that, "Shoppers who buy diapers for the first time at a Tesco store can expect to receive coupons by mail for baby wipes, toys -- and beer. Tesco's analysis showed that new fathers tend to buy more beer because they are home with the baby and can't go to the pub." The worldwide grocer has seen phenomenal returns on their customer analytics efforts. Beyond the beer and diapers example, they "also found that regular buyers of Andrex Aloe Vera [toilet paper] were also big buyers of skin-care products." These insights are important to Tesco for a variety of reasons, but it is important to realize though, how complex it is to yield such insights from data.

Recommendation Complexity


The most intuitive way to make recommendations is to suggest what other people have bought. For example, if buyers of organic cage free eggs usually buy organic milk, grocers will encourage this pattern in customers not yet buying organic milk. Given that there are 25,000 items in the average supermarket though, there are 625,000,000 (25,000 x 25,000) pairs of items to evaluate. We might say that computers can handle these computations easily, until we recognize that customers may be more interested in 3 item groupings... of which there are 15,625,000,000,000. Fortunately there are mathematical methods that can distill all of these combinations into the most interesting ones to marketers.

Clustering


Tesco also assigns each customer a profile using 'clustering' software. "Each product is scored on 50 dimensions such as price and the size of the package. The computer looks for customers whose shopping baskets have similar combinations of scores. [Tesco] classified shoppers in six segments. The 'Finer Foods' segment, for example, is made up of affluent, time strapped shoppers who buy upscale products. 'Traditional' shoppers are homemakers with time to buy ingredients and cook a meal."  This information can later be used when conducting focus groups, to plan new store locations or for targeting new product roll-outs to specific customer segments. Tesco even used it to identify customers that weren't spending their entire shopping budgets at the store. "Clubcard records showed shoppers at a small store... weren't buying full meals. Many people in the town have South Asian or Arab roots." It was subsequently confirmed by focus groups that South Asian families were buying oil and rice in bulk, as well as spices and Asian brands at speciality grocers.

Defensive Marketing


Clustering proved valuable after Wal-mart encroached on Tesco's geographic locations.
"Tesco has used its knowledge of shoppers to fight Wal-mart's core appeal: low prices. After Wal-Mart bought Asda, Tesco searched its database and singled out shoppers who buy the cheapest available item. They were the most likely to be tempted... Tesco figured. Tesco then identified 300 items that these price-sensitive shoppers bought regularly... Tesco lowered the price[s]... As a result, shoppers didn't defect to Wal-mart..." While this article doesn't say it explicitly, the existence of a perfectly price sensitive customer profile is widely accepted (reference The Loyalty Effect). I'm confident that this customer group is one of the six customer profiles cited above, and that clustering thereby contributed directly to the defensive marketing effort.

Value Quantification


  • "In the year that Wal-mart acquired Asda, Tesco's sales jumped 17% to $79 billion and net income rose 17% to $2.96 billion."
  • "the Clubcard has helped boost Tesco's market share in groceries to 31%, nearly double the 16% held by Wal-Mart's Asda chain..."
  • Stickiness: Customers who appreciate the tailored recommendations would be loathe to leave. After all, I'm very unlikely to leave Netflix now that I've invested the effort to rate 1,200 movies because it improves their movie recommendations.
  • If you have your customers' purchasing histories and mailing addresses (which is a sign-up requirement for the loyalty card) you can purchase their demographic information from Acxiom or Dun and Bradstreet. This allows grocers to locate future stores more profitably by identifying demographically similar neighborhoods.
  • Tesco sells much of their purchasing data to CPG companies such as Procter and Gamble, Coca-Cola and Kimberly Clark... some of whom contract with Tesco for direct marketing campaigns to reach customers interested in their new niche products.
  • The purchasing process is simplified for the customer because new products of interest are automatically brought to their attention... adding value through familiarity.
  • Demographic data merged with shopping patterns enables upsell opportunities, for example by adding wine to their stores, or a new line of spices to appeal to specific ethnic groups.
Commentary:
It is worth elaborating on Tesco's retention of price-sensitive customers, even with no profit margin. I believe this is an intelligent decision for four reasons: Tesco retains any incremental discretionary spend by these customers (they would not possess this call option otherwise);  these customers have no variable costs beyond COGS; Tesco can hope these customers will become profitable as their incomes rise; the zero-sum perspective that by harming Wal-mart's profit margins it will encourage them to close.

Wal-mart argues in this article that they can gain the same perspective without a loyalty program by using focus groups and checkout records. I think this assertion is dependent on Wal-mart's target market though... customers with little discretionary spend. These customers will be the least responsive customer segment to upsell opportunities, impulse buys, higher margin products such as pre-cooked meals and wine, and therefore personalization of the experience offers little investment return to the grocer.

"Stores of Knowledge: No. 1 Retailer in Britain Uses 'Clubcard' to Thwart Wal-mart -- Data From Loyalty Program Help Tesco Tailor Products As it Resists U.S. Invader --- Ms. Fiala Changes Detergent" Wall Street Journal June 6th, 2006