"We Know What You Ought To Be Watching This Summer"

The Wall Street Journal article, “We Know What You Ought To Be Watching This Summer,” explains new software that lets e-marketers follow up with consumers based on the attributes of the products they purchase. The software was developed by ChoiceStream Inc., a Cambridge, MA firm.

Previously, marketers like Blockbuster Inc. relied on the obvious, recommending a second horror flick to a customer who previously rented a horror flick, the article notes. But with ChoiceStream the company digs deeper, looking at such features as the complexity of a movie’s plot and whether the movie relied more on plot or characters.

The software then combines what it finds with older data to come up with a recommendation for the customer. The article points out that retailers know improved recommendations keep customers online longer.

ChoiceStream’s CEO believes marketers will appreciate a relationship tool that’s modeled after mom-and-pop stores, where operators get to know customers on a personal basis. The CEO calls it “the most important way of building loyalty.”

The article goes on to explain how the ChoiceStream software found that a Blockbuster customer, who is an adjunct professor of marketing, liked “edgy, urban movies.” Combining this finding with knowledge about his previous purchases of a musical and a documentary, the software came up with a recommendation for a movie called “Rize.”

The article mentions two other companies offering second generation software that makes recommendations. It notes that eBay is working on a way to help buyers find the things they “had no idea” they were looking for, which analysts feel will help the online auction company increase revenue. EBay uses a product offered by a company called Stumble Upon, that looks at customer browsing history.

Overstock.com considers customers’ tastes, then sends them emails. A customer who prefers sleek styling in jewelry would not get an email about “ornate or chunky jewelry.” CEO Patrick Byrne also mentions, as noted above, the benefits in getting to know the customer on a personal basis. He compares it to a Nordstrom salesperson being able to recommend shoes to a customers she’d sold dresses.

The article points out older software grouped customers according to similar products purchased, with a system known as “collaborative filtering,” but that the system requires huge amounts of data and fails for new products.

It adds that basing recommendations purely on past purchases doesn’t work for products purchased as a gift, something Amazon.com discovered from customers who complained. But Netflix’s customers still appreciate being grouped with other customers according to the way they review movies. Pandora.com uses an approach similar to ChoiceStream’s. It looks for musical attributes among 400 it tracks, including the singer’s type of voice. Another music provider, eMusic.com uses ChoiceStream.

Attorney Linus Kafka is not a big fan of the new software. Because he values privacy he calls the software “creepy” and tries to sabotage it by ordering silly things, which to me seems like an expensive way to prove a point.

I enjoyed the article and hope you do, also. I can appreciate the possibilities of more sophisticated ways of getting to know consumers.

I thought mentioning Nordstrom in the article was interesting. Even though the company is legendary for providing personalized service, I came across an article in the early 1990s with the headline, “When Great Service Isn’t Enough.” The clothing retailer went into a slump after losing track of what its customers wanted. I believe it continued to stock formal wear after women started dressing more casual for work. Nordstrom has since recovered, and has a successful Internet presence.
link to article