Case Study: Peruvian Connection Weaves a Stronger Personalization Strategy
The early mantra of online marketing simply was, "build it, and they will come." With thousands of shopping opportunities jostling for attention in today's Web 2.0 world, direct marketers must employ sophisticated tactics to entice browsers and buyers.
"Web site marketing is two things," says Erik Martinez, Peruvian Connection's director of online marketing. "One, getting the traffic, which is why I associate it more with a retail store than a catalog. And, two, once they're in the store, what do we do with them ... what to present, how to present it effectively, and in our case where products are so unique, how do we expose people to the types of products they're most interested in efficiently? Personalization technology allows us to merchandise our products better than we can ourselves [in an online environment]."
Peruvian Connection sells women's and men's one-of-a-kind knitwear along with other apparel, accessories and home décor that have international flair and attention to detail. Until recently, the company relied mostly on gut instinct for cross-sells and upsells on its product detail and shopping basket pages, displaying items that its merchandisers had styled together for grouped product shots (e.g., an outfit that consists of sweater, skirt, belt and shoes). "[They] made sense as products that would sell, but they didn't always sell. People are interested in what they're interested in," Martinez explains.
To get a better handle on what customers and prospects are interested in, Peruvian Connection turned to Redwood City, Calif. personalization technology firm MyBuys, a partner firm to Petaluma, Calif.-based MarketLive, which provides Peruvian Connection's e-commerce platform. The personalization technology analyzes the marketer's customer profiles, transactional data, real-time browsing behavior and other relevant data inputs to deliver targeted product recommendations.
Using a phased approach—which includes keeping a control group that receives manually selected cross-sells and upsells for testing purposes—Peruvian Connection has started with personalized recommendations on product detail and shopping basket pages, moved on to e-mail product alerts, is getting ready to launch dynamic landing pages for search-driven traffic, and has plans for better merchandising transactional e-mail communications.
And it's seeing traction for the earliest rollouts of its efforts; for example, the product detail and shopping basket pages are generating a 25 percent lift in average order value as well as an 8 percent lift in conversions, representing approximately 33 percent improvement in revenue. "As traffic has slowed down due to the economy and some of the timing of our other marketing efforts, this approach is turning into a real revenue-saving and -generating operation, which is kind of fun," says Martinez.
The e-mail product alerts, which involve dynamic merchandising via templated weekly e-mails, leverage customer profile and purchase history to determine who gets contacted and which merchandise gets promoted—e.g., something left in the shopping basket, new products, recent markdowns, etc. So far, Martinez is seeing about a 30 percent increase in productivity for the personalized efforts versus Peruvian's standard e-mail campaign, and he considers the result a good return on spend.
Soon to be deployed, the dynamic landing pages represent an opportunity for Martinez to reduce the site's bounce rates and increase conversions. The approach is to better serve search users' goals by delivering more relevant landing pages based on the search ads and Peruvian Connection's data for which products convert best in each product category for each shopping path. For example, Martinez explains, he could promote a page of the company's best-selling dresses to search users who come to the site off the search term "dresses," but personalization technology can do a better job of serving the most attractive dress products to individual visitors based on deeper criteria.
"It's an art as opposed to a science. This brings some science to the process, while we'll have some art involved with some messaging elements that we control. But the product selection presented should be controlled by an algorithmic- and profile-based tool behind it," Martinez says.
Finally, the static banner of merchandise selections featured at the bottom of transactional e-mails, like the welcome bouncebacks that e-mail registrants receive, will get a targeting upgrade via tests scheduled for the next couple months. While Peruvian Connection doesn't have profile histories on prospects, it does have the purchase histories and best-selling histories on current customers; MyBuys will leverage these insights to determine the appropriate mix of product recommendations. What's even more exciting, says Martinez, is that the personalization technology is "a learning system. As it gains more data, it will target results more efficiently over time, such as customers coming in from this kind of e-mail typically buy a cap-sleeve T-shirt or some accessories."
In all, the effect Peruvian Connection is shooting for is to bring some of the advantages of a retail shopping experience to its site. "I really view the site like a retail site," Martinez explains, "and a well-run retail store gets remerchandised and repositioned frequently ... What a tool like MyBuys brings is some of that intelligence [of visual merchandisers] without adding the extra burden of people [you can't afford to hire] to the process." And the result goes beyond more targeted product offerings, he notes, to the incremental value of conversions.