In my last blog post, I discussed the recent Gartner report “Best Practices in Strategically Combining Search, Content Analytics and E-Commerce.” One of the most important e-commerce search best practices that analysts Whit Andrews and Gene Alvarez emphasize is the ability to “Offer effective definition-matching and handling of ambiguity in Query terms.” Let’s take a closer look at what this means, and how it applies to your search environment.
Effective Definition Matching
The Gartner reports talks about how a truly effective e-commerce search environment must understand the “language variations that are specific to what’s being sold and the audience to whom it’s being sold.” This really boils down to two items a search engine must be able to do:
- For each term in a search string, understand what that value represents – an attribute, product name, product category, etc. – and allow each column to have different relevancies.
- The ability to process search strings of different complexities as entire entities and understand how the individual terms relate in order to return the most accurate results.
This is the essence of natural language. A natural language engine will process a complete search phrase, break it down linguistically and understand the full meaning of the request – NOT just what individual terms mean. In this way, a natural language engine such as EasyAsk can fully support the specific “language of the site” and allow visitors to “speak” to the site in that language via the search engine.
With natural language processing, you can be assured that not only will simple searches – “blue shirts” – be processed effective, but so will complex ones – “women’s blue short sleeve shirts under $50.” You can fulfill this e-commerce search best practice with the most effective definition matching possible.
Ambiguity can come in many different forms. It can come from mistakes or typos. It can come from simple language variations such as different tenses. Or it can come from a visitor’s lack of knowledge of the products – asking for “purple blouses” when none are available on the site. To help you fulfill this aspect of the Gartner best practices, your search engine needs to give you the following:
- Spell correction – your search engine needs to provide automatic spell correction. Anticipating and pre-coding every potential misspelling of each term on your Website is a time consuming task. Who wants to do that?
- Stemming – Your search engine needs to automatically support the different tenses, plurals and other variances of terms. Once again, why should you need to have the time consuming task of entering every potential variance of each term?
- Relaxation – this concept allows the search engine to drop part of a search term if no specific products exist in order to make sure some products are returned. Seeing some products is always better than seeing NO products. With relaxation, a search for “black levi jeans” will still return Levi jeans, even if there are none in black. You search engine needs to have automatic support for this capability.
All of these characteristics will help you virtually eliminate the dreaded “no results” page and dramatically enhance the customer experience by always returning products to the visitor, even when there is some degree of ambiguity.
What if your “site language” is more complex than standard terms? What if your site has a number of acronyms and industry terms? What if you have cryptic model numbers that customers need to use to find parts or products?
To fulfill this last requirement, your search engine needs to make it easy to add synonyms, custom search terms and rules. Once again this is where natural language engines help you implement best practices.
With natural language, you easily specify additional search terms and rules in – well, natural language. You simply type in terms of any level of complexity and associate those with the existing terms or products in your catalog by simply pointing and clicking.
To read more on these capabilities, please download our white paper, “The ABCs of E-Commerce Search: A Guide to Essential E-Commerce Search Features.” In Part 3 of our blog post series, we’ll look at best practices in Search Analytics and Merchandising.