eCommerce Retailers need to be strategic when delivering products that match their customer’s search criteria, if they want to make the sale. Customers are savvy, if they search for a product and it does not appear on the first page of results, they are likely to abandon your site and move on to another. That’s lost revenue.
Site search is getting better these days, and with that development comes the assumption that the most “relevant” results should always show up in the first page of the results display. While that seems like a pretty reasonable expectation, there is actually more to it than meets the eye.
“Relevancy” is actually a loaded term that can have differing meanings depending on how the term is used, and who is using it. EasyAsk understands that there is a difference between “search relevancy” and “business relevancy,” and allows the merchant to control the results display based on the importance of both of these considerations.
Consider this example: A customer of a large online retailer is looking for a television. She types “television” into the search field and sees results that include TVs, but also television remote controls, television antennas, and other television-related products. Those results all contain the keyword she entered and are therefore “search relevant.” But are these results displayed optimally from a “business relevancy” perspective?
It’s a common mistake for a search admin to treat this kind of situation as a search issue and attempt to improve the results by restricting the definition of “television” to return only actual televisions and not accessories. The problem with this approach is twofold: It now makes it harder for the site search function to find accessories for a “TV remote” search, and it doesn’t allow the customer to further navigate to the TV accessories after viewing the actual televisions.
EasyAsk’s technology delivers a different solution. With EasyAsk, the retailer doesn’t need to eliminate results to present the most “relevant” options first. EasyAsk allows customization of how results are displayed through simple “business rules” that weight search results. Using plain English, the retailer can simply type in rules that will prioritize how the results of the search are displayed.
This gives the retailer complete control and solves a problem common to many site search engines. It also has dramatic implications for merchandising, since items can be prioritized by brand, price, whether they are on sale, etc. We’ll explore that avenue further in a future blog.
But back to relevancy: In many cases it’s very difficult for site search technology to know how business relevant a product is to the search request. A related item such as accessory may be categorized in the same place as the looked-for item, and have many of the same words in the product title, but may actually be quite different from the searched-for item.
The site search is doing its job when it finds every product that corresponds to the searched phrase in the correct product category. Now it’s up to the retailer and the display technology to show that information in such a way that the buyer and seller sees what is appropriate in the top results.
To illustrate with another example, let’s say a customer searches for a gas grill. Good site search technology should find all of the products related to gas grills — including the grills themselves as well as accessories such as grill covers and grill cooking implements. A merchandiser with a powerful tool like EasyAsk, however, can control how the search results are displayed with a simple business rule that prioritizes non-accessories when showing equally search relevant products. In this way, a single business rule can have the HDTVs and the gas grills show first for their related searches while also including their equally “search relevant” accessories further down in the results.
The topic of relevancy in search results is more complex than it initially seems. But with EasyAsk, there is also a simple solution that serves the needs of both the retailer and consumer.