The O’Neill Clothing Company is using technology to predict exactly what its shoppers want to buy. And it’s working.

Since implementing technology from vendor Reflektion that personalizes and displays search results in real time, the conversion rate has jumped from 2.7% in 2013 to 4.4% year over year, says CEO Daniel Neukomm. “That’s a substantial jump,” he says. “We high five each other if we can make a .25% change in one direction for new tools or A/B testing.”

The Reflektion technology involves two aspects: a personalized site search and a real time visual search. The personalized search analyzes a shoppers past engagement with the site, including what they have searched for, their clicks, views, and purchases. That data is paired with algorithms that personalize search results for each shopper. The results page shows items that best fit the term searched and that the algorithm has decided the shopper is most likely to buy. Visual search also uses data to personalize search results and begins displaying product images immediately as shoppers begin to type.

“The better at guessing what our want a system is, the more likely you are to buy what it guesses you want,” he says.

Reflektion released the new tools last week, though several clients, including O’Neill, have been running the technology for several months. Those clients have seen increases in site search use in addition to jumps in conversion rates and revenue. “Using Reflektion’s Personalized Site Search, many of our clients’ site search usage has jumped from 6% to more than 30%,” says Sean Moran, CEO of Reflektion. “This is based on repeat visitors finding that search has become a highly efficient way to find the things they are most interested in, and subsequently, conversion rates and revenue increase quickly as well. Increased use of search has tremendous impact on our clients’ bottom lines because it is a higher-converting source, subsequently dramatically impacting revenue.”