It’s easy to think of the search box as an employee. At many online retailers, it services millions of customers a month. But along with direct customer interactions, the search box increasingly serves as a resource to customer service personnel. So, how much does the search box cost an organization each time it fails to effectively comprehend or misinterprets the request of a customer or employee? How can the search box be trained?
Online retailers have begun maximizing search effectiveness and establishing competitive advantages by building structured data. This allows search engines to function like knowledgeable employees. Although technology is a critical enabler of product search effectiveness, technology alone cannot effectively route customers to products without structured product search data.
Below, we explore six primary data strategies that market-leading online retailers are currently employing to drive conversion rates, customer search satisfaction, and, ultimately, online sales.
1. Noun Modifier Pairs
Noun Modifier Pairs are simple, structured statements that reflect the gist of “what a product is,” and as a collective form, it’s the most granular product classification system available to an organization. They often correspond to specific statements that a customer might use when attempting to locate a product in a store. Examples of Noun Modifier Pairs include “Combination Wrench,” “Cordless Drill,” and “Running Shoes.”
Though conceptually simple, Noun Modifier Pair identification and normalization are a complex data challenge for organizations with a large number of unique products. Rules for the creation of these data elements must be applied consistently across the product portfolio and incorporate knowledge of customer search habits.
2. Search Term Synonyms
What happens when a customer makes a search misspelling? Existing technologies exhibit varying degrees of efficiency handling customer misspellings (“florescent” vs. “fluorescent”), spacing variations (“hole saw” vs. “holesaw”), and pluralization (“wrench” vs. “wrenches”). However, they consistently fail to identify search synonyms (“Fuzz Buster” vs. “Radar Detector”).
There is no logic that will identify “fuzz buster” as a “radar detector.” So, the search algorithm must be told that these terms are equivalent. In our experience, the number of unique customer search terms in a given year is often 10 to 20 times the number of unique products offered by an online retailer, meaning that, on average, customers use 10 to 20 different search terms to locate each product. This presents a data challenge to online retailers that cannot be met with technology alone. Online retailers serious about driving product search relevancy and conversion rates actively maintain thousands of search synonyms that are leveraged directly by their site search algorithm.
3. Search Term Routing Maps
When a customer at a store asks for “screws,” a reasonable employee instinctively knows that the customer is looking for screws sold individually or in packs, and is not concerned with the thousands of items in the store that contain screws.
In order to approximate this experience online, frequent search terms often route directly to relevant Noun Modifier Pairs. In order to build these maps, leading online retailers are creating thousands of search term routing “maps” that bypass logic and route frequent search terms directly to relevant products identified by a human employee.
When Routing Maps are used in conjunction with synonym tables, online retailers often achieve dramatic increases in customer search satisfaction and conversion rates. Conversion rates can be tied directly to online revenue as well.
4. Manufacturer Part Number Cross-References
Customers routinely search for items that online retailers do not carry. Invariably, these searches end in null or irrelevant search results. Manufacturer part number cross-references can help online retailers route many of these searches to reasonable substitutes that may be of interest to the customer. For example, if a customer is looking for at 6-foot ladder from a manufacturer that a retailer does not carry, they may be interested in the 6-foot ladder that the retailer does carry.
Like synonyms and routing maps above, the impact of these data initiatives can be tied directly to online conversion rates, and thus directly to online revenue.
5. Product Gap Logs
Many customer searches are null or irrelevant for a very good reason: the product or product type is not offered by the retailer they are searching with and no reasonable substitute is available. Synonym tables and manufacturer part number cross-references cannot fix this issue. However, some online retailers recognize that these searches represent vital business intelligence, and have begun capturing structured information about products customers are searching for that they do not offer.
Product Gap Logs are often built in conjunction with processes used to identify product synonyms and manufacturer part number cross-references. They capture essential information about the product gap that is used to effectively route qualified product line expansion leads to designated Product Managers. Revenue resulting from these initiatives can be tracked over time.
6. Normalized Product Attributes
Online retailers often source similar products from multiple manufacturers, who provide differing product information. This becomes an information organization challenge for retailers and affects their ability to assist customers in “drilling down” within search results.
The most frequent customer search terms are often “broad strokes” that can refer to a large number of items (e.g. “wrench” or “chairs”). Even when a retailer effectively routes these searches, customers are often left to sort through long lists of potentially relevant products. Structured product attributes allow customers to effectively drill down and isolate the exact product they are looking for. They also assist in providing meaningful “Product Compare” views.
When an online retailer only allows customers to drill down using brand names and price ranges, it’s usually because they have not effectively structured and normalized product attributes across manufacturers.
Each of the techniques above is currently being leveraged by online retailers to drive online sales and steal market share from competitors. Due to the number of products and product types offered by online retailers today, these initiatives require significant investment in time and resources, including rigorous planning, program management, and measurement of ROI.
By embracing these strategies early, nimble, middle-tier distributors and retailers have been able to stay one step ahead of larger competitors.