Imported beer brand tricks Facebook's search algorithm
Estrella Jalisco, a beer imported from Mexico by Anheuser-Busch InBev, says it figured out a way to erase some of the vitriol hurled at Mexicans on Facebook.
Previously if some Facebook users typed "Mexicans are" into the social network's search bar, it would auto-fill with negative words, including "lazy," "loud," "animals," and "ugly," according to the brand. But the brew, with help from its agency, David, says it has "tricked" Facebook's search algorithm into producing positive auto-filled terms by flooding Facebook with positive content about Mexicans. Indeed, when tested on Thursday by Ad Age, the search "Mexicans are" auto-filled with phrases like "caring," "great" and "passionate."
The brand and agency lined up a team of Mexican-American artists to create Facebook pages, groups and external websites that were shared on Facebook as articles. When Ad Age performed the "Mexicans are" search on Thursday, a top result was a page including positive images and phrases about Mexicans. The page also included links to external web sites like one called mexicansarepassionate.com.
"We used real people to create content and engage with the content. We did not use any bots," says Yonathan Bendesky, senior brand director for Anheuser-Busch InBev's Mexican imports division. "It really didn't take that much to make a positive change."
Estrella and David created a campaign video, called "Share for Good," (shown above) that describes the project.
Facebook was not involved in the campaign. So-called search predictions (the keywords that are auto-filled as users type into the search box) are determined by a range of factors, including social connections on Facebook, such as friends, groups, pages and events, according to a Facebook spokeswoman. The overall popularity of search terms is also used.
Sun Komen, technical director at Unicorns & Unicorns, a creative studio that worked on the campaign, said that the brand team's testing revealed that "the more people are engaging and liking and commenting on the content that was shared by our team, the more it would suggest that is maybe it is what you are looking for."