NEW YORK • More big companies - PepsiCo, Walmart and Starbucks - are yanking advertisements on YouTube, reflecting their lack of confidence in Google's promise to prevent their marketing campaigns from appearing with vile videos.
The suspensions, confirmed by the companies last Friday, came after the Wall Street Journal reported their brands were placed by Google's automated programs on videos containing racist content.
In addition to the YouTube suspensions, Walmart, PepsiCo and several other firms also said they would stop buying ads placed by Google on more than two million other third-party websites, according to the Associated Press.
The controversy erupted earlier this month after London-based Times newspaper reported that some ads were running with YouTube videos that promoted terrorism or anti- Semitism.
The revelation drove the British government and The Guardian newspaper to take down their ads from the video site.
The snub would potentially cost Google, which in 2006 announced its acquisition of the popular video-sharing platform for US$1.65 billion, and YouTube hundreds of millions of dollars in lost business.
The boycott subsequently spread like wildfire across the Atlantic, as major advertisers in the United States - including wireless carriers AT&T and Verizon Communications - announced they would stop spending on YouTube.
The snub would potentially cost Google, which in 2006 announced its acquisition of the popular video-sharing platform for US$1.65 billion, and YouTube hundreds of millions of dollars in lost business. Last week, US$26 billion (S$36.4 billion) was knocked off parent company Alphabet Inc's market value.
Google attempted to curb the controversy last Monday, pledging publicly to roll out new controls for marketers.
In a memo sent to partners later in the week, the Internet search leader described more detailed changes, including a new video verification process, long sought by advertisers, and a staff hotline dedicated to brand safety.
Another feature Google promised will use machine learning, a type of artificial intelligence, to flag suspect videos. The new approach would now yank ads if offensive language appeared on a T-shirt in a video, for instance, something that did not happen before, according to the memo.
Google aims to implement most of the changes by today, according to the memo, a copy of which was obtained by Bloomberg News. A Google spokesman declined to comment on Friday.
Even if the company meets that deadline, it may struggle to solve the issue.
While television companies have almost total control over what appears on a given channel, creating a safe space for brands, YouTube opens itself up to anyone who wants to post a video.
Advertisers often buy ads across the whole site, or large groups of popular videos, instead of buying ads for a specific channel. The company has safeguards to block offensive content, but the volume of video being uploaded is too great to identify every infringing video.
YouTube warned as much in its memo. Forthcoming changes "should give (advertisers) confidence that their ads will not appear against inappropriate content", the memo read, "albeit with the volume of content involved, this can never be 100 per cent guaranteed".
Mr Paul Verna, an analyst with EMarketer, thinks "what they have done clearly is not enough".
"They must sit down with advertisers and literally show them how they are changing these algorithms," he said.