The Twitter thread claimed that 99.99% of its content is free from hate speech. And over 99% of measured ad impressions were found to be placed adjacent to content considered safe based on the GARM brand safety floor.

They’re nice PR stats, but the question is, does the platform favor controversial content?

Avi Ben-Zvi, GM of North America, Winclap

Still, reports indicate that Twitter’s ad sales plunged 59%, totaling $88 million, during the five weeks from April 1 to the first week of May, compared to the previous year.

“They’re nice PR stats, but the question is, does the platform favor controversial content?” said Ben-Zvi. “For the other 0.1%, what is the virality behind that content?”

Benarroch emphasized that Twitter’s enforcement policy is centered on minimizing the reach of hateful content. While acknowledging the presence of hate speech on the platform, Twitter aims to limit its reach and visibility. According to Benarroch, Twitter’s tracking data indicates that impressions of hate are trending at 0.01% of total impressions.

“What this means is that the visibility of this content is significantly lower compared to other non-hateful, non-toxic content,” he said.

Meanwhile, the platform introduced its first iteration of Adjacency Controls on Dec. 16, 2022 to give advertisers some control over ad placements. It has reported a 99% efficacy rate in diverting ads from undesirable content. The platform said more than 1,500 advertisers are utilizing these controls to safeguard their brands on the platform.

However, according to Ben-Zvi, Adjacency Controls still have room for improvement, such as adding more visibility and acknowledging instances where ads have been withheld from appearing alongside incorrect or potentially hateful speech.

“Showcasing that to advertisers is going to be important because it will reinvigorate their confidence in the platform,” he added.