Business

Struggling to Recruit, Police Turn to Targeted Ads

The pandemic and the George Floyd protests have made recruiters’ jobs tougher. Now they’re tapping the behavioral profiling power of social media.

Your favorite social media platform likely knows more about you than your parents do. Our clicks, likes, and follows reveal patterns that sophisticated algorithms turn into behavioral profiles revealing our political beliefs, sexualities, ethnicities, even our health.

Now police recruiters are tapping these insights to find more job candidates online. Recruiters say their jobs have gotten more difficult in 2021, because of the pandemic and the nationwide uprisings following George Floyd’s murder. WIRED also spoke with digital advertising firms that are working with police and the military for online campaigns to boost recruitment, sometimes relying on the same behavioral profiling tools that platforms use to boost user activity.

“Historically, the majority of our recruitment efforts have been in-person canvasing where we actually go out to schools or trade shows, or meet with organizations,” explains Captain Aaron McCraney, who leads the Recruitment and Employment Division of the Los Angeles Police Department.

McCraney says the LAPD began using the digital marketing firm Sensis in the months before the pandemic. The initial focus was on diversity: The LAPD is struggling to hit its targets for recruiting women, Black, and Asian American applicants.

That can pose a problem for traditional online advertising because employers, including police, can’t target ads at racial or ethnic groups, or prevent other groups from seeing the ad. McCraney says the LAPD traditionally works with certain social organizations—the NAACP for example—to help reach targeted groups. But the pandemic ended almost all offline events, meaning McCraney’s team had to find more women and applicants of color without actually targeting women or people of color. He says the ads have helped.

“Traditional recruiting isn’t working,” says Emma Mae, a marketing specialist for PoliceApp, an online recruitment agency that works with more than 700 police departments in the US. Among other things, PoliceApp creates advertising campaigns and helps applicants through the pipeline. Recently, police departments have come to PoliceApp with interrelated issues: Recruitment is down, while attrition of new hires is up.

This is where the behavioral and psychosocial targeting honed by social media platforms comes in. The LAPD is one of many police departments recruiting by targeting ads based on personality, not identity.

Police agencies want job ads to make the position look benevolent and community-oriented, explains Dallas Thompson, an account director with Sensis. The ads reflect (and hopefully, attract) officers that are service oriented and less money driven, who understand bias, and who have high risk tolerance. Sensis cross-references survey data with the look-alike audiences on social media platforms to identify the characteristics that police agencies say make an ideal candidate: respect for authority, awareness of social biases, interest in service, and a willingness to compromise social life for their career.

As unexpected as the alliance between ad tech and policing is, the tech itself is very well suited for organizing users based on their personality. Social media platforms invest enormous resources into tracking users’ behavior (both onsite and off) and noting what users respond to. They use that information to infer users’ interests and personality, creating the familiar feedback loop that drives millions of people to apps like YouTube and Facebook.

Recruiters design ads that reflect these values and put them online. Wendy Koslicki, an assistant professor of criminal justice and criminology at Ball State University, studied hundreds of hours of police recruitment videos. She says police are fine-tuning ads to display “guardian” imagery. To work around the restrictions on targeting demographics, agencies include women and people of color in their videos, she says.

The videos, she explains, deemphasize weapons and rarely show police making arrests or riding in squad cars. Instead, they stress community work, with images of officers interacting with juveniles at community events, patrolling on foot, and giving speeches in classrooms. Koslicki says the videos often include “statements such as ‘We are a community-oriented department,’ or ‘We value working with diverse communities, we value having officers live in the communities that they work in.’”

Ad tech firms, too, are leaning more heavily into alternatives to demographic targeting amid scrutiny from regulators and the operators of app stores, most notably Apple. That helps when they take on recruiting contracts.

“I think soon companies are going to have no choice,” says Zack Rosenberg, CEO and cofounder of Catapult X, an ad tech startup specializing in video. In Apple’s latest version of iOS, for example, app makers have to ask users for permission to track them on other apps and websites.

Catapult X recently worked with the Secret Service to increase recruitment. Rosenberg says the company’s software uses computer vision, audio recognition, and sentiment analysis to analyze online videos and identify the videos—down to specific scenes—that make sense for a specific ad. The software may recommend a Kleenex ad if it detects an emotional moment in a scene, or, in the case of the Secret Service, may place a recruitment ad in a video that features a soldier returning home from deployment to his family.

The technique shifts the focus from demographics to a user’s emotional motivations. Rosenberg said the most successful military recruitment ads tend to be tied to sports and gaming videos.


More Great WIRED Stories

Products You May Like

Articles You May Like

We Asked AI to Take Us On a Tour of Our Cities. It Was Chaos
OpenAI Slashes the Cost of Using Its AI With a ‘Mini’ Model
Donald Trump and Silicon Valley’s Billionaire Elegy
Waymo Is Suing People Who Allegedly Smashed and Slashed Its Robotaxis
The AI-Powered Future of Coding Is Near

Leave a Reply