(Page 3 of 4)
“Demonstrating how frequently business rules are executed allows retailers to identify what behavior led to the rule being executed or not,” explains IBM’s Frazier. “Automatically updating content in site search based on this information cuts down latency in modifying the site search catalog.”
IBM’s site search engine looks for clues within the site search queries and customer demographics to spot data holes. If a retailer’s customer base is skewed toward women, but still attracts male shoppers, IBM’s site search engine, which is built on the platform used by site search provider iPhrase, will weigh results for men’s version of apparel, for instance, based on the percentage breakdown. IBM, which lists Neiman Marcus, Staples and Restoration Hardware as customers, acquired iPhrase last year.
IBM’s site search engine also makes it possible for retailers to segment the information returned from a site search based on customer loyalty. Retailers can segment customers into tiers based on revenues generated, such as gold, silver, and bronze, and tailor special offers and information contained within the search results based on the attributes of those customer segments.
“Retailers want to group items based on customer segments, not necessarily speed and feed,” says IBM’s Frazier. “It is important to work closely with a retailer’s e-commerce team so that the data model is reflected in the site search experience.”
Doing so plays especially well with corporate clients, which are always looking to leverage their buying power. “Site search needs to be able to consult the rules around each product query and the customer base to deliver a relevant, personalized result,” continues Frazier. “Knowing what a customer buys, how often, and in what quantities boosts the relevance of the search results. Search is no longer a black box algorithm that is difficult to change.”
Such intuitiveness creates a 360-degree view of the information in the retailer’s database, which can then present the customer with a multi-channel view of the products for which they are searching. Fast Search and Transfer allows retailers to cross check site search queries across the supply chain to identify product gluts. Once identified, retailers can offer customers incentives to purchase these items at a discounted price as part of the site search, even if it means heading to a local store to do so. The concept is especially useful for excess inventory of higher ticket items. By enacting such a strategy, retailers can boost margins and the number of units per sale, as well as conversion rates.
When customers shop a web site without using search, conversion rates average between 3% and 4%, compared to 12% for those that use site search, according to Fast Search and Transfer’s Baker. Much of the increase is attributable to the information customers receive about a product or category. “Why serve up just any product if site search can help clear inventory or boost margins, especially when Internet retailing is getting so competitive?” he asks.
Empowering the call center
The intuitiveness that the new breed of site search applications brings to multi-channel retailing is not being limited to a retailer’s web site. Increasingly, retailers are applying site search capabilities to call centers to further empower service representatives.
Endeca has extended its site search capabilities to several retail clients. Service representatives can enter keywords or terms based on their conversation with the customers and pull up information that can guide the customer through the web site to the product or information they seek. “It’s a better use of the data and execution of the retailer’s merchandising strategy by extending personalization into another department,” says Endeca’s Wolf, who adds 1-800-Flowers.com has adopted the concept.
While some users get intimidated at the thought of creating custom search results that can be applied to call centers as well as the web site, Endeca plans to begin creating business rules that deliver customized user experiences based on the keywords entered, how the customer came to a specific department, and their user profile. “It is the way to go when creating a segmented merchandising strategy,” adds Wolf.
IBM is also making its site search technology available to call centers as a way to increase the velocity with which merchandising can influence the online shopping experience without having to rely on intervention from the IT department. Staples and Radio Shack, for example, have applied the technology to its customer service center to help customers find specific promotions online faster.
“Now customers are using the call center to tap knowledge across the enterprise,” says IBM’s Frazier. “Consumers like to push the browsing experience to other areas of the retailer’s business, which is why we do a lot of integration to the user’s CRM application.”
Empowering the store
Another potential use for site search is to make it available in stores. Mercado’s Surles envisions a day when retailers will place kiosks throughout their stores where customers can perform site search before shopping. The experience would be the same as on the web site, that is, delivering relevant products and related information, as well as where the item can be found in the store. Customers can also consult the kiosk to locate out of stock items on the web site or in another store. “The key to successful site search across a retailer’s entire business is using it to keep sales rising quarterly without incurring high costs or expending a lot of resources,” says Surles.
Despite the advances in site search, vendors caution that it is not a panacea for fixing ailing sales. While software vendors can improve performance, it is still up to the retailer to use sound logic when crafting the business rules that dictate site search behavior and how it meshes with their merchandising strategy, according to Thunderstone’s Howitt.
Thunderstone’s engine supports the SQL programming language for customizing site search because it is well known by e-commerce application developers, and lends itself well to incorporating structured logic with search results.