February 1, 2007, 12:00 AM

Treasure hunt

Site search and merchandising technology incorporating more analytics and help from product reviews and content is refining merchandising.

Prior to June 2006, finding the right item at DelightfulDeliveries.com, an online retailer of gift baskets, could be daunting. A search for cookies, for example, would return 39 pages with more than 460 products.

But all that changed when Delightful Deliveries Inc. began using software that melds site search and merchandising functions. In addition to giving shoppers the ability to search by product category, price and occasion, the software enables the retailer to link paid search and e-mail marketing campaigns and Google AdWords to customized landing pages. So a shopper who performs a Google search on cookie bouquets, for example, instead of landing on the retailer’s home page goes directly

to a customized landing page featuring a seasonal selection of cookie bouquets.

“Depending upon what the customer types into the search bar,” explains CEO Eric Lituchy, “besides just coming back with the relevant products we do merchandising specific to that search.”

Since implementing the site search and merchandising application from Mercado Software Inc., last year conversion rates consistently have been 25% higher than corresponding months in 2005, Lituchy says. Orders per visit have increased 50% and there has been a 100% increase in units bought per visit, he adds.


Like Delightful Deliveries, many e-retailers are using site search tools to gather information about customers to boost sales through more targeted merchandising. Now they’re adapting site search to tap into non-traditional search categories, such as customer product reviews and informational content, and to work with web analytics programs.

Using site search to personalize merchandising efforts is not new. In fact, the term ‘searchandising’ was coined several years ago to describe the concept. Many site search tools identify and fix failed searches, produce search results and navigation options most likely to lead to sales, and push products based on a retailer’s other business objectives, such as increasing sales of high-margin items or reducing inventory of overstock merchandise.

But new site search products are incorporating more analytics, such as visitor behavior and clickstreams, to further refine the merchandising process, says Susan Aldrich, senior vice president at Patricia Seybold Group, a research and consulting firm that specializes in customer-centric I.T. strategy.

The push to make site search more analytical comes at a time when many Internet retailers are trying to dig deeper to get a feel for which promotions and products generate the best return on investment. “People need to know more than just how many searches they had,” Aldrich says. “Can you imagine running Macys.com based on how many times people searched for perfume? How would that tell you what to promote? How does that tell you how successful your promotions are?”

In earlier forms of searchandising, a search engine would make cross-sell or upsell recommendations based on what a customer bought on an earlier visit. However, that strategy has flaws, contends Phil Braden, director of customer-facing solutions at Endeca Technologies Inc., a vendor of site search and merchandising software. For example, a customer who bought a book on Yoga might be shown Yoga merchandise every time he comes back to the site, regardless of what he is looking for on the return visit. “It may turn out that he bought that book for his sister and he doesn’t like to do Yoga at all,” Braden says.

Site search can fine-tune that merchandising process by making recommendations based on the customer’s current search query or category navigation choice. “Search and dynamic navigation elements are a give and take, a call and response between the user and the information systems,” says Joe Lichtman, director of retail product management at Fast Search and Transfer, a search technology vendor. “With each search, they’re telling you more about themselves and what they’re interested in.”

Feeding time

To get a clearer picture of what customers are searching for, Petco Animal Supplies Inc. expanded the attribute fields in its Petco.com database from five to 58 fields when it implemented Endeca’s site search tool, says John Lazarchic, vice president of e-commerce at Petco.

“We use site search to drive any aspect of what a customer wants to use to shop for the product, including traditional shopping navigation and search refinements-color, size, flavor and brand,” Lazarchic says. “But we also added some unique items to that customer refinement process-the ability to either sort or refine items by customer ratings and reviews.”

Petco feeds average ratings of its products generated by the Bazaarvoice Inc. customer review application into the Endeca engine to enable customers to sort products by customer ratings or reviews. The technology also has enabled an automated process that upgrades left-side navigation based on top-rated items. As products become rated better or worse, they automatically fall into or out of that top-rated category, Lazarchic explains.

“It really gives the customer the power to decide what way they want to find something,” he says. “As a retailer, you tend to get stuck in your own paradigm of what you call things or how you think things should be categorized. You quickly learn that customers don’t think of things in the same way you do.”

At Delightful Deliveries, Mercado staff culled data from the retailer’s catalog when setting up the search engine database. The data includes item number, product name, brand and price. But it added new information created by the retailer, including which gifts might be appropriate for certain occasions or recipients.

By adding even one attribute to a search query, a shopper can narrow results significantly, says Sue Chapman, senior manage of merchandising solutions at Mercado. For example, by entering a price range while searching for cookies, the results page returns 219 items, rather than the 464 products listed when simply searching for cookies. By searching for cookies to serve at a baby shower, results narrow even further. Delightful Deliveries customers also can use more complex search term combinations-for example, cookies, Mrs. Fields, baby shower-to more quickly find the product they’re seeking, Chapman adds.

Similarly, Macys.com uses dynamic browse refinements that enable shoppers to narrow down choices based on specific features, such as capacity and speed for blenders or warranty length and special features for watches. Macy’s, which uses Mercado software, also has implemented a feature that enables shoppers to compare the products in the shoppers’ refined search results.

comments powered by Disqus




From The IR Blog


J.T. Compeau / E-Commerce

How Walmart is getting its Oscars debut right

Consumers talking about the Oscars on social media are also engaging with Wal-Mart, data shows.


Mike Cassidy / E-Commerce

5 e-retail planning tips for holiday 2017

Monday’s turn out to be prime shopping days during the holiday season.

Research Guides