June 7, 2005

The Dark Side of Web One-to-One Marketing

By Evan Schuman, Ziff Davis Internet

Opinion: A University of Pennsylvania study raises questions about how online retailers are identifying potential customers and how they are using that information.

It has been said that a university's role is not necessarily to preach the right answers, but to teach students how to ask the right questions. To the extent that's true, a recent University of Pennsylvania report seems to be properly performing its role.

The report itself was hard to ignore; it reported that 64 percent of consumers "do not know that it is legal for an online store to charge different people different prices during the same time of day, a practice used by some companies that base their prices on what they know about a customer's buying habits."

The upshot of the Penn report—that consumers are remarkably clueless about e-commerce practices—was of little journalistic interest. I have given up having any faith in the American public ever since it gave "Laverne and Shirley" the highest TV ratings for a disturbing percentage of its run. So, rampant consumer ignorance on Web issues seemed to be a big Dog Bites Man story.

Consumers want to know that they are being treated not only fairly, but equally. The suggestion that the existence of a cookie—showing perhaps showing a prior visit to a price comparison site or even a rival vendor's site—could impact pricing is alarming.

Ironically, that capability—if used—would be little more than what a good sales rep should do in the field: Know the customer and make pricing proposals based on everything that is known.

But the underlying suggestion that there are tons of e-retailers out there reading cookies and changing prices in real time is intriguing.

Despite some widely circulated consumer media reports to the contrary, the study didn't actually establish that any retailers are setting prices on an individual consumer level based on cookies, other than mentioning the five-year-old story about Amazon.com having experimented with different prices for people based on what cookies were found on their machine.

Even the Amazon reference is thin, because five years is a long time for the Web and everyone was experimenting with everything back then.

The examples cited in the report are much more innocuous, including the practice of grocery stores offering instant coupons worth 75 cents off Brand X yogurt to customers of Brand Z yogurt.

Technically, that is offering one customer a product at a lower price than another, but the transparency at the checkout counter seems to prevent any strong resentment. It's being accepted as customized coupons.

But the concept of cookie-influenced pricing struck a nerve because it touches on a long-term retail industry effort to discourage customers who are low-margin, not-so-affectionately dubbed bottom feeders.

It was also using cookies—which most consumers just barely understand—as a way of punishing price-comparers. As a wannabe-bottom-feeder myself, the suggestion didn't sit well.

Are e-tailers actually doing this on any kind of widespread scale?

Joseph Turow, the University of Pennsylvania communications professor who managed the report, conceded that there is no proof of any retailer doing it, but he said that he strongly believes it's happening.

"Companies today collect more information digitally about people than they ever had before," he said. Are they using it to customize pricing on an individual consumer basis? "All you have to do is read between the lines."

Paula Rosenblum, the director of retail research for the Aberdeen Group, flatly said that no U.S. retailers—of any note—is scanning for cookies and changing prices on-the-fly for individual customers.

She did add, however that, in Japan, retailers have been experimenting with prices that change on the store shelves based on various factors including inventory and time of day, but that's not customized to the consumer.

Personally, I'd say the answer is somewhat in between. The answer is either "No" or "They are, but they're keeping it really quiet."

For "no" to be the answer, we would have to believe that tracking and analysis software doesn't yet work that smoothly. I have no problem finding that credible.

For the "cloak-and-dagger" scenario to be true, we'd have to believe that the retail IT people—who tend to move from company to company—and the various software vendors involved have all managed to keep this capability quiet.

I'm reminded of a favorite line from "The West Wing," when White House Press Secretary C.J. Cregg complained about trying to contain leaks: "There is no group of people this large in the world that can keep a secret. I find it comforting. It's how I know for sure the government isn't covering up aliens in New Mexico."

The caveat, though, is "widespread." There are a lot of sites out there with some very creative and persistent programmers behind them.

Are there tiny pockets of e-commerce sites that are trying to use cookies to set on-the-spot pricing?

Undoubtedly, just as there are some e-commerce sites somewhere that are undoubtedly phishing for credit card numbers, too.

Given that an e-commerce site today can be some guy and his cat, generalizations should be used carefully.

Rosenblum added that this does bring up an issue being debated in e-commerce circles.

More and more retail chains are seriously looking at regional pricing, where California stores may charge more for the same items than, for example, Iowa stores.

But if a retail chain sets prices by region, how does it do that on its Web site? Is it handled in the same way as state sales tax, where a consumer is charged a different amount depending on the ZIP code where the item will be shipped?

What if item isn't shipped, such as downloadable software? Will cookies—or an IP address—be used to try and identify a consumer's true geography?

Rosenblum said that it's certainly true that retailers—BestBuy is a frequently cited example—are trying to discourage bottom feeders, but it's mostly being done by de-emphasizing—or simply ceasing to carry—products that seem to attract the bottom feeders.

In other words, the reverse of standard stocking mentality: Find out what they want and then make sure we get rid of that.

The vendor perspective is somewhat more tempered. Scott Friend, the president and co-founder of ProfitLogic, a major software house specializing in retail pricing strategy, said that none of his clients are using his software—to his knowledge—on Web sites to deliver that kind of individualized pricing, but that the capability is pretty much there.

The typical changes are sometimes weekly or price changes for different styles of products based on popularity, inventory or other factors. "But it certainly doesn't change per customer," Friend said.

He sees the potential for such uses in the future to be less disagreeable.

"I don't see anything unethical about" using cookies to set pricing, although he quickly added, "invading someone's privacy is certainly not within the realm" of his software and stressed that it was a matter of "tailoring the assortment to each and every customer."

Friend talked about the various product comparison sites. "If the consumer is using the tools, why not the retailer?" he asked. Retailers "should have the ability to filter out" customers they don't want, such as bottom-feeders.

Although the report is likely to stir industry debate, analyst Rosenblum had few nice things to say about the University of Pennsylvania's research, but her unintentionally nastiest attack was when she referred to the school as Penn State. At Ivy Leaguer U. of P., them's fightin' words.