• by Laurie Sullivan  @lauriesullivan, February 7, 2019

Purging millions of accounts may have helped Twitter redeem its reputation when it comes to providing a safe haven for brands to interact with consumers.

The results -- from a study fielded by GumGum, a marketing analytics firm, released Thursday -- suggest that 45% of the small sampling of 274 marketers participating cite Twitter as the most brand-safe social platform in 2018, up from 1% in 2017.

Renewed faith in Twitter by marketers shows in its fourth-quarter 2018 earnings report released Thursday. Twitter posted fourth-quarter 2018 revenue gains Thursday, up 24% to $909 million. Total advertising revenue came in at 23% to $791 million, about $425 million in U.S. advertising and $366 million from international.

In GumGum’s study, Facebook came in at a distant second with 26%, while YouTube took the third spot with 11%.

Just one year ago, 34% said Facebook was the riskiest platform for brands. At that time, publisher sites placed second with 27%, and YouTube came in third 15%.

GumGum surveyed brands, agencies and publishers in November 2018. The findings shows that marketers are at least somewhat pleased with the steps social media platforms have taken to address brand safety since 2017.

Despite all the work, vulnerabilities remain. Some have negative side effects, such as the overuse of blacklisting websites that Ben Plomion, GumGum CMO, says could prevent marketers from reaching their target audience.  

Some 62% of marketers still relieved heavily on black listing websites in 2018, up from 50% in 2017. But 62% of marketers using this tactic said their ability to target their audience declined in 2018, up from 30% in 2017.

The perception of brand safety has changed. In 2017, 90% of marketers were concerned about brand safety. In 2018, 60% said they still felt the same.

Plomion attributes the 10% decline to the perception that social media platforms improved safety standards. Take, for example, Facebook’s and Twitter’s decision to delete thousands of fake accounts tied to Iran, Russia, and Venezuela. These events happened as recently as January 2019.

Other tactics cited by marketers as successful in protecting brands include natural language context detection, image recognition, direct relationships with publishers, keyword detection, geotargeting, whitelists, and ads.txt, which Google spearheaded.

Thirty-five percent of respondents said they use natural language context detection as their most effective tool, up from 18% a year ago. About 31% use image recognition, up from 13% a year ago.