Mixing It Up

With its selection of high-end designers mixed with lesser-known labels, Intermix's recent performance is a bright spot on an otherwise bleak retail horizon. Its blend of established and emerging designers offered at varying price points – plus its strategy of pairing pricey pieces with moderate ones – fills a niche the company's founders identified in 1993.

Their fashion strategy -- think Sharon Stone pairing black Gap T-shirt with Valentino skirt at the Oscars -- clicked and Intermix quickly expanded from a single Manhattan boutique to two dozen stores in additional markets such as Atlanta, Boston, Chicago, Dallas, Las Vegas, Los Angeles, Miami and Washington, D.C.

Customers can expect to find the designs of Stella McCartney, Diane von Furstenberg, Alexander McQueen and Michael Kors alongside up-and-comers like Mary-Kate and Ashley Olsen's brand, Elizabeth and James.

Don McNichol, Intermix's vice president of direct, was brought aboard in 2005 to launch IntermixOnline.com as part of the company's multi-channel initiative. Prior to that, Intermix had captured 1,200 e-mail addresses from customers who subscribed to its e-mail news feature. (In fact, McNichol says, many Intermix shoppers look forward to Friday nights when they receive Intermix's weekly e-mail, browse the website with a glass of wine, discuss new items with friends and then shop online or in-store over the weekend.)

McNichol wanted to strategically target Intermix's online and in-store customers based on RFM, which stands for recency, frequency and margin. Intermix was already working with Novato, Calif.-based iPost for targeted e-mail solutions, and McNichol turned to iPost to help extract more value and greater margins from each customer.

Intermix "wanted to apply some of the usability research we learned through the old heat map sensoring, eye movements and the like to a new way of looking at merchandising online and e-mails," he says, "and to design some performance metrics around that in order to understand what is most relevant to the customer."

Craig Kerr, vice president of marketing for iPost, says the company's Autotarget was designed to help medium-sized online retailers and bricks-and-mortar stores engage customers through targeted e-mail campaigns. Autotarget's segmentation and predictive analytics help companies identify less-engaged customers (who subsequently receive fewer e-mail offers) and better and best customers, who might receive more relevant e-mail offers based upon their unique profiles.

Using Autotarget's RFM analysis, Intermix could automatically place its customers into micro-segments, called cells, based on past behavior including click rate, open rate and purchasing.

Identifying customer discounts
Before adopting Autotarget, Intermix's open rate hovered around 20 percent and its click-through rate was about 35 percent. In addition to boosting these figures, McNichol wanted to be better able to identify who to give a discount to, how much of a discount was necessary to motivate certain customers and how long Intermix could wait before offering discounts. He also wondered how often he could e-mail customers before being spit out by a spam filter.

With Autotarget, McNichol is able to draw behavioral data from two Intermix sources -- e-mail responses from iPost's e-mail tracking database and Intermix's own purchasing behavior data. Moreover, a medium-sized retailer that doesn't employ a large analytics staff "could have actionable data immediately, rather than waiting days or weeks," Kerr says.

Intermix uses Autotarget's predictive analytics data to send targeted test mailings to customers in specific cells, which are divided into three groups:

• VIPs represent 20 percent of the customer base, have a higher amount of disposable income, will shop for the latest trends and demonstrate low price sensitivity.

• Sale Shoppers represent 40 percent of the customer base, are aspirational customers desirous of Intermix's fashions but demonstrate high price sensitivity.

• Brand Shoppers constitute 40 percent of the customer base, shop specific designers, like a "good deal" and demonstrate only moderate price sensitivity.

McNichol conducted A/B tests to determine the most effective offer for each segment. (Intermix continues the Friday e-mails and uses that as the day to update the site, from the home page to landing pages.)

Different strokes, different folks
For the VIP group, Intermix offered a gift card with purchase rather than a straightforward discount. E-mail offers to this segment might also include invitations to special private events like designer receptions or invitations to be among the first to see new lines and collections.

The Sale Shoppers segment received a 30 percent-off offer, a steep discount meant to motivate them to action.

For the Brand Shoppers, the offer was a 10 percent to 15 percent discount along with information on specific brands.

Using Autotarget, McNichol found ways to maximize e-mail communication for those customers who wanted more contact while minimizing discounts and increasing discount lag time, thereby shortening the discounting period. Following the test period, Intermix saw a 90 percent increase in its e-mail open rate (now 38 percent) and a 46 percent rise in click-throughs (now 51 percent). At the same time, Intermix's opt-out rate declined from 1.2 percent to 0.19 percent.