JPMorgan Chase develops algorithm for brand-safe ad placement on YouTube
- JPMorgan Chase has created its own proprietary algorithm that will ensure its YouTube ads will not appear next to questionable or unsavory content, according to a Business Insider report. The major U.S. bank pulled its YouTube ads in March after they appeared with inappropriate content and has expressed frustration with YouTube's slow response to the problem.
- The algorithm, developed in-house by JPMorgan's programmatic and media-buying teams, plugs into YouTube's API and uses 17 filters to segment safe channels to host its ads. Filters include a channel's total video count, subscriber counts, topics, native language and comments on videos.
- Among the 5 million YouTube channels, the bank has pared down its safe list to 3,000. After rolling out the tool in October and testing it, resulting in a 99.9% success rate, the company has returned to advertising on the video platform.
Moves like JPMorgan Chase's suggest YouTube hasn't been successful so far in rebuilding trust with brands following the platform's issues around brand safety last year even after putting in place a series of new measures to better ensure ads don't appear next to inappropriate content. Given YouTube's significant reach, marketers may have no choice but to take matters into their own hands like JPMorgan Chase has done.
"The attention of protecting a brand has to fall on the actual people within the brand itself," Aaron Smolick, executive director of paid-media analytics and optimization at JPMorgan Chase, told Business Insider. The exec also said the issue is really about appropriateness rather than safety, with each brand needing to decide for itself what is appropriate.
According to the Business Insider report, JPMorgan Chase will perform routine checks and continue to refine its tool so that its ads remains safely positioned. The bank is one of the marketers at the forefront of trying assert more control over where ads appear online. Last spring, the marketer significantly cut back the number of websites serving its ads programmatically with little deterioration in performance metrics.
JPMorgan Chase is among the long list of brands — along with McDonald's, L'Oreal and Audi — that pulled their advertising after seeing their ads pop up next to YouTube videos of hate speech, extremist groups and other unsavory content. The bank was also not the only brand to express frustration with YouTube's slow handling of the problem and take matters into their own hands. Several marketers have brought their advertising in-house and began reducing the number of channels where they place their ads.
YouTube initially felt marketers' pain over where ads were ending up in the form of a 5% decline in advertisers in April. However, the platform soon bounced back with a 50% increase in Google Preferred advertisers. In August, YouTube announced that it would use improved machine learning technology to better spot and remove certain videos.
Brand safety was an ongoing concern for marketers in 2017, as many realized they really have no control over where ads are actually placed, especially when they use programmatic buying strategies. About half of the marketers in a study by the Society for New Communications Research of the Conference Board didn't know where their ads were running online, but also weren't sure how to adjust their strategies for safer placements. In the SNCR study, 88% of marketers said having ads placed next to false or sensationalized information could harm the trust they'd built with consumers.
And, they were mostly right. More than one-third of consumers see ads as endorsements of the content they're placed around, even when that content is offensive or inappropriate, according to a Survata survey commissioned by AdAge. However, 55.1% said it didn't change their opinion about the brand.