Youtube Adpocalypse: the product of AI shortcomings, a divisive political climate, and a hyper targeted ad strategy.

The Boycott Begins

In late March, major brands began removing their ads campaigns from all of Google’s non-search advertising platforms following the London Times’ investigation revealing that many of the companies’ ads had appeared tacked to videos promoting terrorism and propagating hate speech.

Screen Shot 2017-04-21 at 4.19.27 PM.png

Guardian, a British newspaper being advertised over a Jihadist video

Over the course of the next week, more than 250 companies, including accounts as big as AT&T, PepsiCo, Nestle, Starbucks, Verizon, and many more, had frozen their YouTube ad accounts. Given the fact that Youtube’s business is based almost solely on ad revenue, the controversy packed a major punch for Google (Alphabet), whose stock was downgraded from “buy” to “hold” buy multiple firms in the days following the Times’ release. There have been various estimates as to how much revenue will be lost as a result of the boycott, but the average floats around $750 million. Despite Google’s huge losses, it’s actually the creators that are feeling the harshest effects, so much so that they’ve named the entire fiasco “Adpocalypse.”

How did this happen?

It’s not like Google hand-picked controversial videos to tag ads to–they actually don’t hand pick any ad placements. With over 65 years worth of video being uploaded to youtube every day, the company is forced to rely on algorithms to match advertisements to videos. Between the time a user clicks on a video and the ad begins playing, an algorithm is used to match one of thousands of ads to that video at that time for that viewer in order to optimize probability of engagement and conversion. Included in this algorithm is targetable categories laid out by the company, demographic and viewership data about the individual viewer, information about the ad’s previous engagement history, and information about the video’s content, tags, ratings, and comments in order to categorize the video itself. The video below explains the advertisement matching process in more detail.

Google’s current ad-matching algorithm system is one of the best in the industry–it uses machine learning and artificial intelligence combined with data mining to evaluate trends associated with the content, the ad, and the viewer to make the best suggestion possible in just milliseconds. The problem with this, however, is that algorithms cannot pick up on every sentiment in a video, and has yet to be able to weed out certain types of content all together, the way traditional advertisers would be able to when pairing their advertisements to a television network or specific show. Online, video is uploaded so quickly that there is no time for person to person interaction, no time for an explanation of content or assurance of value alignment. And given today’s intense political polarity, brands are even more sensitive to the political sentiments they’re giving off, making the uncertain nature of Google’s algorithms even less attractive. Furthermore, Google’s European boss recently criticized many of their brand partners for not fully taking advantage of the controls they offer.

When AI fails

Not only did Google and content creators lose significant sums of money throughout the boycott, but the incident highlights a larger problem with digital advertising. No matter how well they are programmed, machines and algorithms cannot yet match the emotion sensing and judgment capabilities of humans. They’re efficient, but they have their downfalls, and the recent Adpocalypse shows just how detrimental those downfalls can be. Google’s team, in response to the boycott, has widened the definitions of “offensive/hateful content,” and is also looking to increase required channel standards in order to have ads placed on a creator’s videos. While google has been attempting to decrease the variability of their ad placement, there is a limit on non-human sorting abilities, a reality we are going to have to get used to in such a digitized world. Furthermore, tightening algorithm criterion can have its own consequences on the creators supporting the content the ads are tied to.

The end of a content democracy?

Given a (warranted) lack of trust in automatic ad placements, major brands may be more inclined to focus their online/content advertisements on more “human-based” approaches, like influencer marketing wherein brands company with online personalities for product promotion and placement deals. This would ensure companies could have a greater input in the values they’re promoting and greater control over how their brand is portrayed. Such advertisements already contribute to a major percentage of influencers’ income, however, this is only the case for the top tier influencers with follower counts in the millions.

In their post-boycott security update plan, Google noted that they will limit ad placements on ‘legitimate’ channels within the Youtube Partner Program, essentially suggesting that the ability to monetize through ads will now also be limited to creators with larger followings. This is something Youtube is already doing: currently, they have certain channels with reputable histories whitelisted–these channels are given more successful ads and are given ads more frequently. Channels that are known to be consistently dealing with problematic topics or have a history of being reported are blacklisted, and receive no ads whatsoever. Channels not on either of these lists are at the mercy of the algorithm. Essentially, Youtube is looking to expand both of those lists and give a higher proportion of ads to those on the whitelist, thereby sharply decreasing the revenue that Youtubers not on either list can make.

Youtubers with content focused on comedy, politics, and even things like video games have already claimed to have ads removed from their videos. Now, in order to have ads roll before a video, a creator must be mainstream, extremely popular, and completely non-controversial. RobinHood1If this trend continues, the democratic nature that Youtube has long been known for will likely fall. The popular will get more popular, the rich richer, and the smaller creators creating niche content or in the development process will only have their growth stunted, disincentivizing smaller creators. Those creators who are reliant on ad monetization may now face difficulties in planning their future revenue models, and will likely reconsider the risk they took in dedicating more time to Youtube over another career path. Without the small income supplement from video monetization, smaller creators may no longer find it worthwhile to continue making videos on a regular basis or continuing to truly pursue Youtube as a career/secondary job. 
Though I do believe that Youtube will be able to recover from Adpocalypse and almost all of the companies partaking in the boycott will begin advertising again shortly, I think this incident highlights a lot of the shortcomings of the current online advertising model and will result in at least a subtle shift in how brands continue to leverage an increasingly important online market, and will definitely affect the politics of the Youtube community.


  1. JoshLArtman · ·

    This is a great post! Two weeks ago I watched a video where a few Youtubers explained how they’ve been personally affected by the “Adpocalypse,” but I didn’t understand it was such a complicated and widereaching problem. According to that video, any content that is even slightly “not family friendly” can no longer display ads – this broad category apparently includes videos about guns and even the LGBT community. This whole debacle shows that content creators shouldn’t place all their eggs in one basket, because the platform they depend on for their livelihood can change the way it functions at any time. Perhaps more Youtubers will look to alternative funding platforms like Patreon in order to combat the Adpocalypse. Really liked the topic of this article and all of the hyperlinks, thanks for sharing!

  2. Nice post. I’ve also wondered when looking at less objectionable but still disturbing content (i.e. on CNN or something), whether companies would be happy with their ads being served up beside that content.

  3. Thanks for posting! I had heard about the issues with ads being placed on ISIS propaganda videos on YouTube and the backlash from brands, but I didn’t know how widespread the issue was. This reminds me of brands removing their content from sites like Breitbart and even shows on TV like The O’Reilly Factor. It will be interesting to see if Google can improve their algorithm in the long run to help curb some of the concerns brands have.

%d bloggers like this: