Rick Griskie has a simple way of describing the new “natural language” search functionality on ADESA.com that the company says makes finding vehicles faster, easier and more robust. He describes it as a “Google-like search.”
Griskie, who is ADESA’s chief information officer, explains that today, most people find vehicles through features. The searcher first enters the manufacturer such as Ford. Then the system asks the searcher to choose from a list of models, such as an F-150. Then, the user must drill down through additional filters, which Griskie describes as “quite cumbersome.”
“Natural language search, quite frankly, is what Google introduced us to when they came on the scene a number of years ago, where instead of using filters and drop downs, you just started typing in the … search box what you’re looking for,” Griskie said in a phone interview Wednesday.
ADESA said in a news release that its new search capability uses what it describes as “the leading cloud-based search technology,” which can return up to 400,000 vehicles in less than two seconds. ADESA said that is six times faster than before.
The natural-language capability is a first for the remarketing industry, ADESA said. The enhancement, paired with ADESA.com’s advanced filter options and canonical database, substantially reduces the time a user spends searching for the right vehicles, the company said, adding that it improves productivity and profitability for dealers of all sizes and sales volumes.
Griskie also explained the canonical data in a simplified way, describing it as a “decoder ring,” and mentioning the usual search through make, model, and trim.
“However the data we receive from the numerous disparate sources of that data [doesn’t] follow the make, model, trim very cleanly,” he said. “So we’ve created a canonical database that is effectively the decoder ring. So for example, in the trim field for many of the data sets that we receive, trim is blank. But we’ve created a proprietary way to ascertain what trim that vehicle really is, and that goes back to our canonical database. And then more people that use our natural language search, the smarter that canonical database becomes based on what people are searching for,” he said.
Griskie added in a news release that “The most crucial activity for dealers sourcing inventory is vehicle search.”
“Our search now rivals anything you’d find in the online shopping space. It works even when customers misspell an entry. Plus, it learns from itself, quickly adapting to help buyers more effectively with repeated use and giving customers more of what they want and less of what they don’t want,” he said.