A Blog by Jonathan Low


Jun 29, 2012

Custom Fit: Etailers Personalize Appeals - Increasing Sales vs Creeping Out Consumers?


Short hand for 'too much information,' and indicative of knowledge perhaps better left unknown. Because it's too personal, too intimate, too, well, creepy.

But this appears to be the way that internet advertising and emailing is headed. Powerful algorithms crunch data on purchases, web site clicks or whatever other information can be gleaned and quantified. The result provides a statistical probability that if you liked X, you may well buy Y.

Netflix, Amazon and other e-retailers use the resulting wisdom to make suggestions for purchases whenever you click on their sites. And that seems to be ok with the majority. Books and movies, whatever.

But we live in a world where living well enough alone is perceived to be a weakness and aggression is the new reticence. Especially when it comes to selling online.

The economy is weak, online sales are still in their relative infancy, many people continue to shop the old fashioned way - in-store, in-person - so the competitive urge is to push as far as one can to land the sale. The evidence suggests that such tactics work up to a point. But there are those who are discovering they have limits, even though they may be perfectly comfortable surrendering all manner of personal data to third parties for the sake of convenience and an occasional discount.

Identifying and establishing those limits is a work in progress. Facebook seems to find new ones every week. Presumption is not a trait generally appreciated in our society. Some of the uses of data may be crafted in ways that offend or violate that precept. For some it may be phrasing, for others, intent and implication.

We will eventually learn more from neuroscience than from ham-handed abuses of societal mores by over-caffeinated Gen Yers, but all data points are useful - for the hunted as well as the hunters. JL

Natasha Singer reports in the New York Times:
Amazon may have introduced product recommendations tailored to individual customers years ago, but greater personalization is now fine-tuning the face of e-commerce with bespoke shopping experiences. Retailers look at customers’ locations, past purchases and current online activities to customize content to individuals in real time in an effort to increase sales.
DMITRI SIEGEL, until last year a marketing executive at Urban Outfitters, thought he had hit on a novel idea to personalize the company’s Web site for frequent customers. He would make it easier for female shoppers to peruse women’s apparel and for men to concentrate on men’s clothing by altering the site’s product displays to match a user’s gender.

Unknown to many consumers, a firm called Monetate helps businesses fine-tune their online marketing by analyzing individuals’ locations and behavior.
It seemed like a no-brainer.

“If you could just stop marketing dresses to men, it would be amazing,” Mr. Siegel said last week about his thought process at the time.

With the help of a Web site testing and optimization company called Monetate, Mr. Siegel experimented with gender personalization on the site. But it roundly backfired. It turned out that many female Urban Outfitters customers regularly bought men’s items and they took offense at being subjected to gender-based marketing.

“We saw customer frustration at being targeted outweigh any benefit,” said Mr. Siegel, now the vice president of global e-commerce at Patagonia. “If you got it wrong once, it outweighed getting it right 10 times.”

The customization is often covert. Many people are unaware that sites may show them material or offer deals — like discounts or free shipping — that are different from what their neighbors see. Yet the phenomenon is growing. Half of the largest online retailers in the United States used some personalization techniques last year, compared with about 33 percent the year before, according to Internet Retailer’s Top 500 Guide. And e-tailers are turning to a handful of specialty software companies like PredictiveIntent, RichRelevance, MyBuys and Monetate to help them analyze customer data and segment their audiences for special treatment.

Tailoring sites to users can spur online sales, repeated site visits and in-store sales, says Kurt Heinemann, chief marketing officer of Monetate.

“That helps our clients increase revenue and creates more relevant experiences for customers,” he said. Users don’t have to click 10 times to find products they like, he added, because “the site is bringing things to them.”

But when personalization gets too personal, as Urban Outfitters’ executives learned, it can come too close for many consumers’ comfort. It turns out that hyper-customization may produce reactions similar to the “uncanny valley” effect in robotics in which people find themselves repulsed by humanoids that too closely resemble human beings.

If e-tailers become too familiar with users, they risk alienating them, says Mahender Nathan, the vice president for e-commerce and digital marketing at Godiva, another Monetate client. Personalization, he believes, should adhere to the conventions of in-person conversations.

“In conversation, if you think it’s odd that you know something about someone that they didn’t share with you, don’t use it,” Mr. Nathan said. “What we’re trying to shoot for is friendly, cordial and helpful as opposed to crossing the line and being creepy.”

Godiva, for example, is using the Monetate system to test a promotion tailored to users in warm-weather regions. The promotion guarantees that the chocolatier’s products will arrive in perfect condition no matter the climate.

Monetate, based in Conshohocken, Pa., has designed technology that marketing executives can use to modify or run experiments on their Web sites using natural language, not computer code. With the company’s dashboard, a marketer can instantly customize a site to appeal to, say, customers from Los Angeles, people who are interested in watches, big spenders — or high-end Los Angeles watch collectors — offering those particular audiences $50 off a purchase of $500 or more. The system can compare the actions of shoppers who receive tailored offers to those of people who did not. The dashboard charts the results, displaying the impact of site changes on new customer acquisition, average purchase amount per customer, overall sales and return on investment.

Mr. Heinemann of Monetate says personalization is just the latest incarnation of old-fashioned customer service.

“It’s like a 1950s retailer, where people knew their customers and could steer them to things they knew they would like,” he said. “You perform better by tailoring the experience.”

Founded in 2008, Monetate has a client list including Aéropostale, Best Buy, Patagonia, Petco, QVC, Revolve Clothing and Urban Outfitters. Those companies tend to frequently tweak their sites, he says, running dozens of tests at a time to gauge the impact of seemingly small changes like font size or free shipping on customer behavior and revenue.

TAKE, for example, Revolve Clothing, which sells 400 different brands of designer jeans and other fashion items.

Two years ago, the site introduced a feature, called “My Revolve,” by which customers could select the designers they preferred, creating their own mini-boutiques with their brand choices, says Kobie Fuller, the company’s chief marketing officer. That allows Revolve to send tailored e-mails promoting customers’ preferred brands. But, he says, it is too limiting for the site to simply depend on people to divulge their fashion preferences because only a small percentage of visitors voluntarily share their tastes.

“The next steps really revolve around not actively asking visitors for their likes,” Mr. Fuller said, “but taking the data that Monetate is tracking — on what people are viewing and buying — and showing them images or brands they like.”

Revolve already tailors its offerings based on whether a customer comes to the site after using a search engine or after clicking on a promotional e-mail. It also offers free shipping to users from certain countries. Those kinds of tweaks have resulted in seven-figure increases in annual gross revenue for Revolve, Mr. Fuller says.

Certainly, many people welcome such tailor-made conveniences. As personalization becomes more pervasive, however, it has the potential to invisibly channel consumers who might mistakenly think they are browsing freely, limiting their experience to the views algorithms have predetermined they might like and act upon. Retailers say they are grappling with the question of how to make customization helpful without being invasive or constrictive.

One way to do that, says Mr. Siegel of Patagonia, is to personalize features that add value for consumers. Right now, for example, Patagonia is experimenting with climate and geographic customization, showing waterproof gear to customers in Seattle during rainstorms and surfer boardshorts to users in Southern California. Knowing a customer’s location, he says, enables Patagonia to be more precise about delivery times.

“There’s a lot of hype around personalization. I am deeply skeptical,” Mr. Siegel said. “The big question right now is: ‘Is it right for my business, is it right for my customer?’ ”


Post a Comment