The call for an audit of Facebook’s metrics comes a week after the social network acknowledged inflating its video metrics.
The Facebook Like button and simple questions bring more of personalization into the open.
When a shopper arrives at ShoeDazzle.com one of the rotating central images asks her in bold letters, "What's your style? Sexy? Boho? Classic?" Clicking the link takes her to the site's "Style Survey" to answer such questions as "Which cosmetics are most likely to grace your face?" and "Which diva's style would you dare to wear?" With each answer, indicated by clicking on one of three images, the site zeros in on what shoes she might be interested in buying.
The retailer, which sells subscriptions to receive new footwear each month, uses that data to send customers shoe recommendations. The consumer then selects which items she'd like to receive. Based on how consumers respond to the selections offered, the retailer fine-tunes its algorithm to better predict what its customers might like. The process requires having customers who are willing to interact with the retailer, but that works for a fashion-focused brand like ShoeDazzle, which has nearly 1.1 million Facebook followers, says Hope Neiman, the retailer's chief marketing officer. "We're involved with our fan base and they're involved with us," she says.
And the retailer asks customers to help in making its selections more relevant to each shopper. "It's not just one big black box," Neiman says. "Before the month starts we ask them what they're looking for next month. When they contact our client services team through Facebook or on the phone, they tell them what they're looking for and ultimately that information informs the algorithm." While Nieman would not reveal ShoeDazzle's sales, investors were sufficiently impressed that they put $40 million into ShoeDazzle earlier this year.
Like ShoeDazzle, many online retailers are becoming convinced of the value of personalizing the content they present online shoppers. Half the retailers profiled in the 2011 edition of the Internet Retailer Top 500 Guide employ personalization on their e-commerce sites, up 53.4% from 32.6% a year earlier. Moreover, 61% of retailers say personalization is among the most important merchandising tactics in web retailing, according to the 10th Annual E-tailing Group Merchant Survey, which canvassed 200 retailers of various sizes and selling in a wide array of product categories.
The idea of a retailer letting customers know it's asking for their preferences so it can make better recommendations is hardly new. Netflix Inc., for instance, has for years asked subscribers to rate the movies they viewed and made plain it would use that data to recommend other movies they might like.
But it's relatively new to see that approach show up on sites like ShoeDazzle.com and comparison shopping engine TheFind.com as a way to sell items like shoes and apparel. TheFind uses a variety of techniques to personalize the content it presents, and among the most recent innovations is to rely on a new Internet phenomenon that has millions of consumers broadcasting their product preferences—the increasingly ubiquitous Facebook Like button.
To do that, the site encourages consumers to personalize their shopping experience by logging in to TheFind.com using their Facebook user names and passwords and allowing TheFind to access the stores and brands the consumer and her Facebook friends have tagged as Likes.
The system, which TheFind.com's developers designed with Facebook, then sorts search results based on the brands and stores that a consumer and her Facebook friends Like. When she clicks on an item, the preview product window shows the names and photos of friends in her network that Like the product or brand. The window also displays how many Facebook friends that brand or retailer has on its own Facebook page. The Facebook Like button appears so that the consumer can add herself as a fan of the brand or store.
Even if a consumer is one of the roughly 75% of consumers who do not use their Facebook log-in on the site, the site's default uses the broader Facebook community to help determine the order of its results. Retailers with more Likes generally appear higher in the results. When a consumer mouses over an item, the site displays how many consumers Like the retailer. "There are enough retailers on Facebook that we feel like we can treat it as a store rating system," says Ron Levi, TheFind's vice president of product.
The site also uses a consumer's IP address to present her with localized results when she clicks on the site's Local tab. The tool, which works even if she isn't signed in, is sometimes right, and sometimes it isn't, says Levi. However, having it automatically set up—and having an Edit button that allows the user to change the location—encourages consumers to interact with the site. "We want people to connect with us," he says. "It's a gateway to getting people to share more."
The site also uses a cookie to gather a consumer's previous search results to populate the site's MyFinds tab, which highlights a shopper's past product searches.
Because TheFind searches more than 500,000 stores, having several ways to arrive at relevant results is useful, Levi says. "We want to save consumers the trouble of having to search through 40 pages of results. And we'll take any signals we can to help them do so."
Wisdom of the crowd