Expanding our commitment to brand safety

Twitter’s purpose is to serve the public conversation and we are committed to creating a healthy environment where everyone, including advertisers, can participate and engage safely. We know there is a lot of work to be done to achieve this, and we are actively working with key industry partners to advance brand safety as a central component of the advertising and measurement solutions on Twitter.

Here’s a look at our areas of focus ahead of 2021: 

Doubling down on third-party brand safety measurement solutions.

Today, we're excited to announce that we've selected DoubleVerify (DV) and Integral Ad Science (IAS) to be Twitter's preferred partners for providing independent reporting on the context in which ads appear on Twitter. We see this as an opportunity to build solutions that will give advertisers a better understanding of the types of content that appear adjacent to their ads, helping them make informed decisions to reach their marketing goals. These solutions will complement our existing third-party viewability measurement solutions with DV and IAS.

Over the past five months, we’ve engaged in a rigorous process to select partners from a slate of third-party measurement vendors. We are excited to make this announcement, knowing that there has never been a third-party brand safety solution built to measure for adjacency to brand unsafe content in an environment like Twitter, which is both real-time and dynamic. We look forward to partnering with both DV and IAS to create custom solutions for our unique platform and we intend to start testing solutions in early 2021.

Our commitment to MRC accreditation. 

Last week, we also shared our commitment to undergo the accreditation process across all four of the Media Ratings Council’s offered Accreditation Services: Viewability, Sophisticated Invalid Traffic Filtration, Audience Measurement, and Brand Safety. We will be prioritizing the Brand Safety audit first, and have already started the proposal process for this accreditation. 

We believe that all four accreditations are critical in demonstrating our enduring commitment to transparency and are pleased to be in a position to expand on and sustain this collaborative process. We look forward to working with the MRC and further aligning on scope, sequencing, and timing.

Continued partnership with industry leaders. 

Twitter is an active and founding member of the Global Alliance for Responsible Media (GARM). In conjunction with our ongoing efforts to better align our brand safety solutions with GARM’s principles, we support DV and IAS in their efforts to adapt their current categorization frameworks to better map to the harmful monetized content categories identified by the GARM, the 4A’s APB, and the ANA. We will continue to partner with industry leaders to define and refine the work that is needed to keep platforms safe for advertisers. 

Unlocking research to learn more. 

Twitter’s personalized feed-based environment makes content adjacency inherently unique when considering brand safety questions. It is critical to understand where associations occur, where they don't, and how those associations affect brands as we continue to invest in this area. Early research done by Twitter’s data science teams indicates that on our platform, adjacency between ads and divisive content has not yet been found to affect brand favorability. Currently, we are working on additional research to gain more knowledge and insight around the impact of adjacency on Twitter ads. We look forward to incorporating these findings into our ongoing work with DV and IAS and using it to inform how we build and iterate on our advertising products.

Marching towards 2021. 

We are excited to build on the existing momentum with our brand safety partners and to develop innovative solutions for our customers that help push our industry forward. We will continue to be transparent in all our efforts and keep you updated along the way. Most importantly, we are ready to do the work.

Study Methodology: We analyzed Twitter Brand Survey data from April to July 2020, which included 50,000 respondents who were exposed to 1,200 Promoted Tweets. For the purposes of this study, we used a broad definition of divisive content, inclusive of 11 content categories including Politics, Social Movements, Ongoing News and Protests. For the purposes of this study, adjacency was defined broadly as any of the two Tweets above and the two Tweets below each of the Promoted Tweets and we analyzed results based on content both 1 and 2 spots away from ads. Analysis was performed both in aggregate and specific to each of the 11 categories of divisive content to determine if adjacency to certain types of divisiveness alter people’s opinion of adjacent brands. Results showed no evidence of any effect on brand favorability when ads were adjacent to any of the studied categories of divisive content, regardless of the distance of the adjacency.