Retailers Use Transaction Data to Get Specific in Mail Campaigns

As a way to compete with new selling channels, mom-and-pop retailers and large department stores are putting more marketing muscle into targeted direct marketing campaigns based on robust databases.

Industry experts say retailers are shifting from general media — such as billboards and newspaper advertising — to targeted programs and proactive customer-retention strategies. Retailers have filled recent DMA and NCDM trade-show floors to learn new strategies for competing with the Internet and other channels and to find ways to hold marketers more accountable for returns on their marketing programs.

“There has been a tremendous push in the last year on database marketing and relationship marketing in the retail arena,” said Evangelos Simoudis, vice president of solutions at IBM Global Business Intelligence Solutions, Armonk, NY. “Retailers have been reading about the successes of banks, telcos and insurance companies that are using their databases in a much more integrated manner and trying to find out how they can learn what products they should be selling to specific companies.”

John E. Roberson, senior vice president of marketing and sales at Dynamic Marketing Services Inc., a KnowledgeBase Marketing Co., said, “Retail times are tough. Retailers are trying to show sale increases over the past year and trying to find ways to do things differently to improve these numbers.”

Their customer bases are fragmented and as a result, “you have to go direct. You have to find ways of reaching out to people who are your customers,” Roberson said. “Retailers are trying to develop relationships with their customers instead of just hoping to hit loyal customers.”

To meet those needs, leading database companies such Harte-Hanks Direct Marketing, IBM and Dynamic are offering products and educating retailers. IBM has found that department stores with catalogs are using sophisticated database marketing techniques for their loyalty programs.

“They are starting to apply their databases in ways that will allow them answers to questions such as, 'Which product should I offer to this customer and over what period of time?' This is something they never did,” Simoudis said.

Before, he said, retailers used to look at customers mailing by mailing. They would put out a catalog on a certain date and send it to specific households using data from their less-than-sophisticated databases. Now, they are using their databases to see how many catalogs they have sent to a particular household over a period of 18 months and how often customers buy from those catalogs.

“[These retailers] can use this data to decide whether or not it makes sense to keep sending catalogs. In general, catalogers are measuring the profitability of risk,” Simoudis said.

He said retailers are merging their catalog customer databases with their databases of credit-card holders who pay in-store for merchandise. By merging these databases, “companies can make projections about profitability and risk of the customer, but also understand customer preferences. You can also find out about behavior here, such as how customers pay and what kind of risk profile they have.”

Peter Robson, vice president of database marketing services at Harte-Hanks, Baltimore, said retailers are becoming better at leveraging transactional data — data collected at the store's register that shows who the customer is and what he or she is buying — and linking it to other retail databases of names and addresses of customers gathered from private-label credit cards, bank cards and bridal registries, for example.

He said retailers are storing more than 24 months' worth of transaction-based line-item details, such as what the customer bought, when he or she bought it, what store it was purchased in, who the sales associate was, how it was paid for and if it was on sale or full price. Transactional data that discusses credit activity collected in-store is five to seven times more predictive of future shopping behavior than other attributes, Robson said.

“Customers are finding that what is most indicative of future relationships with their customers is their past credit activity,” Robson said. “It's more important to learn what they bought and when they bought it than what their ages are.”

Liz Claiborne Inc., New York, the women's sportswear manufacturer with more than 100 U.S. stores, understands the importance of past credit activity. On May 2, the company sent its full-price Summer Style minicatalog to more than 100,000 of its best customers at either the 50 Elizabeth stores (for larger women) or the 50 Liz stores that reach a more general audience. Harte-Hanks helped sort out the transaction data to define the audience.

The customers were selected form Liz Claiborne's Preferred Customer File, which contains point-of-sale data. When a customer makes her first purchase, a cashier collects information such as name, address, telephone number and her likes and dislikes. Once the customer is registered, every subsequent transaction is recorded.

At the end of each month, Liz Claiborne's in-house MIS department compiles a tape of all of the transactions that is shipped to Harte-Hanks for cleaning and merge-purge. Here, the data is put onto Harte-Hanks software, which allows Liz Claiborne to build queries based upon customers — such as how many customers spent more than $1,000 or how many responded to a promotion.

Although no information is available yet on the results of the May promotion, a March mailing to best customers garnered above-average response.

“We love doing surveys of our best customers once a month because it allows us to quantify our direct mail efforts,” said Marc Cohen, Liz Claiborne's director of marketing for core retail. “Also, when we do the back-end analysis, we are always looking for build in terms of our ROI.''

With this data, Liz Claiborne can decide whether the mailing programs are successful or whether they should go further into the database.

Transactional data is a key part of Dynamic's Proximity/Propensity Model, which overlays the data with Dynamic's compiled database of 155 million. With this, Dynamic can analyse the retailers' best-customer profiles. This information is overlaid with longitude and latitude plots so that marketers can learn the profile of someone who has the propensity to behave like a best customer and understand how the best customers relate to the proximity of their stores.

“This process allows us to find, for example, a group of customers that spends $1,000 a year with a retailer, responds to six offers a year and lives within two miles of the store,” Roberson said. “Then we can find that another segment spends $500, responds to five offers a year and lives within five miles of the store.”

Retailers, however, still have a ways to go before they reach the standards of their catalog cousins, such as L.L. Bean or Lands' End.

“[While] many retailers certainly talk about these techniques and the analytic activity that gets companies close to customers in a very detailed kind of way,” Robson said, “fewer are integrating it into their overall marketing programs.”

Related Posts