Everyone has had the experience of looking for a product online and getting zero results because the search was too specific—or, inexplicably, getting an almost infinite number of unrelated results. Most ecommerce databases simply weren’t built to be queried in an intuitive way, and their basic relational structures and hierarchies can’t handle the complexities of the modern web. This has a number of consequences for enterprise, including lost sales, brand image problems, and customer frustration. According to Mark Unak, CTO of Codifyd, Inc., many consumers interact with a website about eight times. If they can’t find what they want, they’ll move on, and that spells trouble for companies that need longer sessions to sell products.

Codifyd remedies this problem by employing Marklogic’s schema-agnostic databases along with their own XML-based system that allows for separate, unlimited attributes in metadata. “We understand what a product is through the metadata. And then each product can have an unlimited number of tag value pairs for attributes. So we can understand in the context, if you search for a ¼” slotted magnetic screwdriver, we solve that looking at the metadata rather than looking through the skews.”

Many developers and companies have issues with content management systems, because they weren’t designed to merge data like video, pictures, and rich text. But being able to store, query, and analyze heterogeneous data allows companies to both recommend more related products and upsell to customers based on the precise features they’re looking for, instead of keywords or what other people have viewed. “What we bring to the analytics part of it is the broad product information that can then marry with the customer and really understand buying behavior,” Unak said. “Why you made that buying decision is going to be more exposed. And we as merchandisers can then go back and try to influence your behavior.”

Watch the full interview below, and be sure to check out more of SiliconANGLE and theCUBE’s coverage of MarkLogic World 2015.