Jennifer Zaino
SemanticWeb.com Contributor

There’s a lot of data out there on the Web – too bad it’s still so difficult for the average person to find what they really need.

The problem of improving the user experience and presenting people with the data they’re searching for using intuitive methods will be addressed at the upcoming Web 3.0 conference in a session on Data and User Interfaces. Consider the difficulties of finding data on the deep web, where web pages are dynamically generated based on back-end databases.

“Finding that information is not an easy experience,” says Steve Lavine, CEO of search, discovery and recommendation vendor Transparensee, who will be speaking at the conference. “The user is the person in the worst position to make search-related decisions. They don’t know what is in the database.”

As a result, they’re often not exposed to potential results that actually have applicability to their requirements, even if they’re not exact matches.

Transparensee provides next-generation search and tagging services to content providers, such as online dating sites outfit WorldSingles and nationwide apartment rental search company ForRent.com, where most of the information users search on is stored in databases behind the firewall.

Annoying your end users

Lavine says “the million or none” problem – where users wind up with either too much information from a search or too little, and either way they think the site doesn’t have what they are looking for – is a big issue for sites, because it leads to users abandoning a page without finding what they need and without being converted to a lead or making a purchase. He hears from businesses all the time that it’s great that Google gets users to their site, but then their sites’ poor search defeats expectations with poor data presentation.

“The end user sees a marked difference between the web search experience and the database data mining search experience,” Lavine says. Equally annoying to end users are the results they sometimes come up against thanks to sponsored ads on web sites, where top-of-the-list paid results often have little or nothing in common with the users’ requests.

The approach Transparensee takes to the issue of improving search results for databases of discrete data is to put complex business rules under the hood to accommodate the true preference of the user, no matter what those are, and make searches faster so that users can manipulate results in real time. It’s developed proprietary technology that enables its clients to map their data to deliver search experiences that guide the user in interactively specifying their search requirements, making informed assumptions that expand the search to what they’re also likely to want to see-but don’t necessarily know exist-based on both metadata and keyword requirements. Users then can adjust those assumptions to generate even more precise searches.

For example, a user may be searching on a realtor site for a house in a certain zip code with a certain number of bedrooms at a certain price, and in conjunction with an exact match to those requirements Transparensee also can deliver results for homes in the same zip code at that price with three bedrooms – a close match. Users can adjust that to weight having four bedrooms as being more important than the zip code or price, for example, and re-generate the results to support discovery of homes exactly fitting their room requirements criteria and still close to their other desires. Or perhaps the user wants to see only homes in a certain price range but where listings also contain the word “marble”-Transparensee can accommodate that too.

Semantic smarts come in its categorizing of customers’ tags for a particular dimension or category and building a structure so that its system has an understanding of similarities, closeness or relevance across each category. It not only improves the data results for the type of requests that are better searched by numbers (zip code, price, review ranking, etc.), but can pair that up with the ability to match language terms.

For example, a user searching at a restaurant review site may want to find a place to eat that can offer him the same kind of food he has eaten at another restaurant, and he may want the alternate place to be in the same zip code, offer the same prices, and be similarly ranked, as well. The restaurant he has eaten at before may be categorized as offering “Emilian” food-a particular kind of Italian cuisine-but since the searcher may just want a plate of pasta, Transparensee is smart enough to also deliver results for restaurants categorized in its taxonomy as Italian, Roman, or under other related terms that meet the other requirements, as well.

“The semantic web really is a mapping of terms-you create the relationship and take the onus of searching off the end user and put it on to technology,” says Lavine.

That’s a road to improving customer satisfaction through better interfaces and smarter data presentation. “I think that’s the next stage in this evolution,” he says. “It’s not just saying, ‘Here’s Google and do your best to search,’ but here we can make a lot of assumptions for you, we can learn about you, and give you the information we think is best suited for you. We can do a lot of the work for you and then just offer you choices. That’s a big evolution in the way that search on the web is working.”

Semantic Tech & Business Conference Returns to San Francisco

Semantic Tech & Business Conference returns to San Francisco in June! Join us from June 3-7 for complete coverage of Big Data, Linked Data, Extreme Information Management, and Semantic Web. From breakthrough approaches to solving business problems to the big data implications of fast–evolving technologies, SemTechBiz provides you with an unparalleled interactive experience and delivers tangible business value. We're offering a special early rate when you register by February 17. Sign up now!