In the wake of scandals, platforms are policing audience segmentation

“In the nature of the self-service model, you’re handing [all the targeting] over to someone else so it was important for us, for a number reasons, to be very thoughtful of what segments to work. We want to make sure we’re honoring the trust we feel like we’ve earned from the users, to create those segments in the first place,” said Foursquare’s svp of product, Josh Cohen.

Foursquare’s team eliminated several possible categories such as visitors of gun stores or strip clubs. Rather, Foursquare decided to launch categories that were broader like road trippers and nightlife. Snapchat has similar targeting called “lifestyle categories.” Snapchat currently has 117 of these categories. Foursquare released 450 segments at launch.

Foursquare also decided to eliminate categories based on scale, making sure the segments are large enough so that users cannot be personally identifiable. People’s specific homes are not categories, for example.

Twitter, meanwhile, restricted the ability to target small audiences, buyers said.

“Before, on Twitter, you could do 200 to 300 people. You could be way more targeted, but now it shows up to 1 million. [Twitter] took away the niche factor. Now you’re at the mercy of Twitter’s algorithm,” said Nick Venezia, managing director at Social Outlier.

In its attempts to prevent misuse, Facebook, has also been eliminating niche groups for their audience segmentation. Last August, the company said it was removing more than 5,000 targeting options as part of its efforts to prevent abuse on its ad platform.

“Facebook has made it harder to see smaller audience sizes where the old limit was around 1,000 on some days. The list size column will now say unavailable,” said Duane Brown, founder and head of strategy at Vancouver-based agency Take Some Risk.

But while these platforms are trying to clean up, it’s still possible to game the system.

Facebook’s platform still allows for nefarious ad targeting. Earlier this month, the Los Angeles Times reported that it was still possible for advertisers to target Facebook users who were interested in Nazis.

Advertisers on Twitter, meanwhile, can generate their own tailored audience based on their customer lists of email addresses and phone numbers. One buyer said that if they were granted that ability, they could work around Twitter’s restrictions of not targeting someone’s followers by simply using Twitter’s API to export a user’s follower list and then upload that as a custom audience. It’s easy to see how bad actors could take advantage of this, and it would violate Twitter’s rules.

Meanwhile, Google claims it has been extremely restrictive to prevent nefarious activity. Its rules, which ban targeting to sensitive interest categories such as identity, beliefs and personal hardship have been in place for at least a decade, a Google spokesperson said.

Though, a buyer said it’s still fairly easy to get specific on Google.

“Google is still the wild west. There are segments available at low CPMs for almost anything. For all I know, whatever is in these segments could be BS, but it makes clients feel better thinking they can target a woman who has a size 7 shoe with a low credit score and is in the market for an insurance plan for her house that she hasn’t bought yet,” the buyer said.

Brown said his clients have never been interested in nefarious targeting, but saw how it’s challenging for the platforms to prevent it.

“The challenge is you are asking people to change something but in one part of the code. Are you changing the base code is my question? It’s like focusing on your queen in chess and ignoring the rest of the chessboard, which is your base code. Just trying to band-aid a solution is not going to work,” Brown said.