What COPPA Means for the Future of YouTube Brand Safety
YouTube is no stranger to scandals, but what’s happened in the last few weeks of fall 2019 might spell disaster for everyone’s favorite place to share cat videos.
How bad are we talking?
Thanks to COPPA (The Children’s Online Privacy Protection Act), a law passed in the late 90s, there’s a legitimate possibility that some of the most popular YouTube channels will disappear next year.
Why would a data scientist like me care? Well, it affects all of us in the digital advertising industry and may be a sign of things to come.
Let’s start at the beginning…
How YouTube’s COPPA issues started.
The year was 1998 and the internet was still in its infancy. This was only 6 years after the first ever image was posted to the internet:
Even this early on, it was clear that the internet was here to stay. Anyone could hop on the internet and their data could be collected with impunity. Someone with remarkable foresight must’ve thought “It seems like a bad idea for random websites to collect data on children” and COPPA was born.
COPPA’s purpose is to protect kid’s data online. The entire act is pretty dense, but to summarize: it forbids the collection of data or online tracking of children for advertising. In this case a child is someone under the age of 13.
Your Title Goes Here
The act does allow for the collection of data if a parent/guardian’s permission is obtained, but the requirements for obtaining said permission are so complicated that most companies just opt to not track them at all.
Fast forward 8 years to 2005 and the birth of YouTube. To avoid getting in hot water with the federal government, YouTube needed a way to not break COPPA rules while still collecting data on users older than 13. Their solution?
The honor system.
When signing up for a YouTube account, you have to verify you’re at least 13 years old. COPPA is why.
This may seem kind of ridiculous since a kid could obviously lie and claim to be 13 by just checking a box, but this was considered good enough. Presumably the FTC thought “Well, it’s not perfect, but at least YouTube is trying…”
Your Title Goes Here
I have no idea what the FTC thought back then or thinks now.
Fast forward to 2019. The FTC reviewed YouTube and its compliance with COPPA. It didn’t go so well.
What happened? Did the FTC change their mind? Was there something YouTube’s lawyers missed? Did YouTube get caught collecting data in some sort of Watergate-style bust?
YouTube outed themselves.
Remember how I said kids could easily lie about their age to join YouTube? Well it turns out they did. By the millions. How do I know that? YouTube admitted it.
And they didn’t just admit it. They bragged about. YouTube went around making claims about how many kids use their platform to entice advertisers.
Advertising to kids is big business. I feel dirty just writing that, but it’s true. According to this article from the American Psychological Association, it was about 12 billion dollars annually in 2004. This number could have doubled since then.
The reason for this amount of money is that, according to Digitas, kids and tweens as a demographic have 1.2 TRILLION dollars in spending power (that includes money their parents spend on them, but still).
If you’re approximately as old as the first image on the internet you may remember watching Nickelodeon and Cartoon Network on TV as a kid. These channels made money by selling advertising space for a child audience. Toys, cereal, kids movies; these are big business and they have ads like any other product.
I think it’s obvious that some kids would end up on YouTube despite their terms of service requiring a user to be 13, but YouTube shouldn’t be able to tell if they’re actually kids. So how did this happen? Well, YouTube was somehow able to determine there were kids on the platform despite all accounts having to state they’re over 13 (the precise details of which haven’t been released).
Since YouTube had a valuable audience — children — they decided to capitalize on it. They told advertisers about it and just sort of forgot that it was supposed to be a secret.
It was the equivalent of cheating on a test and then bragging to your friends about it while the teacher is in earshot. Except it was a billion-dollar company and instead of academic fallout, there was financial fallout. A lot of it.
The YouTube COPPA fallout:
Unsurprisingly, YouTube was caught red-handed and forced to pay a settlement: $170 million.
That’s a lot of money, but not for YouTube — whose exact annual revenue is hard to pin point — but let’s take the low end of the estimate from this article and put it at 16 billion. That puts the fine (in the most devastating scenario) to be a little over 1% of their annual revenue.
To put this in perspective, MacKenzie Bezos got approximately $38 billion in stock as part of her recent divorce from Amazon founder Jeff Bezos. A hundred-million-dollar fine just doesn’t hurt like it used to.
That wasn’t the only fallout. As part of the agreement, YouTube will not allow targeted advertising on kids videos. Ads fall into 2 categories: targeted/behavioral and contextual.
I’ll give a brief summary of both types, but if you’d like to read way more about it, I wrote a whole article on the subject: What is Contextual Targeting?
Your Title Goes Here
What is behavioral targeting?
With behavioral targeting, an advertiser can pick who they want to see their ad.
For example, a shoe company might want to be able to target based on gender since (generally) male and female shoes are distinct categories. The exception being Crocs (which are the epitome of fashion regardless of gender) and the Rocky Horror Picture Show (if you don’t understand that reference you’re probably younger than the first picture on the internet).
Behavioral targeting is helpful for advertisers because it allows for greater accuracy in regards to who sees your ad. Consequently, targeted ads are worth more. However, they require data in order to work and COPPA does not permit collecting data on kids.
That means that going forward, kids videos on YouTube will only be able to use contextual targeting (if they use targeting at all). That’s bad news for anyone making kids videos.
Your Title Goes Here
It’s worth taking a step back to remember that YouTube is a bit unique among social media. Channels on YouTube are making money off the ads on their videos. For the larger channels (at least 1 million subscribers) they’re probably making videos as their full time job – or at least primary form of income. High quality videos take time and money to produce. In other words, some of the most popular channels (not just kids channels) will be most affected by changes to YouTube’s algorithm.
A kids channel on YouTube may be looking at a significant loss in revenue by January 2020. You read that right, January 2020 — less than a month away. How significant is hard to say from the outside, but one content creator estimates 80-90% of a YouTuber’s revenue could be affected by these changes.
Admittedly he’s a biased source since he is a content creator on the YouTube platform, but with 12 million subscribers I think it’s safe to say he understands this platform pretty well.
Your Title Goes Here
I watch this channel all the time and his video is what inspired me to write this article. I recommend watching the whole video if you want a content creator’s opinion on the situation.
It gets worse, though. You may assume that YouTube is going to attempt to classify videos using machine learning, but the primary method of classifying videos will be by the creators themselves.
Content creators will have to classify their own videos as either a kid video or not.
You may be wondering why content creators have to identify their own videos. After all, wasn’t YouTube able to figure out that there were millions of kids on the platform despite all users verifying there were at least 13?
That’s true. YouTube could have used any number of methods to estimate their child audience: statistical modeling, machine learning, survey results, you name it. However, YouTube is not using it to provide a guide on what is “kids content.”
Maybe it’s because YouTube’s own method isn’t 100% accurate. Maybe they don’t want to take on that legal liability. Or maybe they were forbidden to give that kind of guidance as part of the FTC settlement. It’s unclear at this point. So, content creators are on their own here.
Alright, we’ve covered a lot here. Let’s recap:
- Content creators will have to label their own videos as being “kids content” or not.
- To comply with the FTC ruling, YouTube will disallow targeted ads on kids content.
- Most of revenue a YouTube channel makes comes from targeted ads.
With all that in mind it seems better to just not label any videos as kids content. But it’s not that easy.
Sharks in the water.
The qualifiers released by the FTC for what constitutes a “kids video” range everywhere from super broad to applicable to nearly every video on YouTube.
Some are obvious, such as the 5th most-watched video on the platform: Baby Shark (4+ billion views). Others are much less clear. Is Minecraft considered for kids? The game is rated 10+, but I certainly know people well into their 20s and 30s who play it.
So, what if you can’t tell the difference? Who cares if a creator accidentally mislabels a few videos?
Well, the FTC cares. And they can hit any mislabeled video with a $42,530 fine per video. Somehow the fine given to YouTube as a platform was barely noticeable, yet the fines that can be given to creators on the platform can be financially devastating.
FTC Fine for Mislabeled Videos: up to $42,530
This leaves many creators in a horrible position. They have two options:
- Label your videos as “for kids” and accept a huge cut to your monthly revenue, OR
- Risk the wrath of the FTC and financial ruin.
A change this drastic would require months or years to prepare for. And it’s coming in a few weeks to all creators on YouTube.
I’d like to make a comparison to really drive home how unfair this situation is:
Imagine this happened to traditional TV. All TV channels, including streaming services like Netflix and Hulu would have to self-identify as “including kids content” or “strictly adult.”
This would mean that channels that include kids content would lose 80% of their revenue starting in 2 weeks, while adult channels could lose 42k per incorrectly labeled show.
I think most would agree that this would be ridiculous, but that’s basically the system YouTube and the FTC are implementing.
It’s easy to dismiss these concerns. After all, no one wants YouTube to be overtaken by catchy-shark-themed-music-videos, but not everyone on the platform deserves this kind of punishment.
What about animators?
What about video game speed-runners?
What about cat videos?
Do these count as kids content and are they at risk? The depressing answer is we don’t know.
Why should advertisers care?
If you’re reading this you clearly care about digital advertising. There’s a good chance you run digital ads and you may even run ads on YouTube. A change this drastic could destroy some of the most popular channels on the platform and could even lead to the decline of the platform itself.
In some ways, YouTube is more comparable to a TV subscription rather than a social media platform. Earlier this year, HBO panicked as the end of Game of Thrones approached. The media giant feared that millions may cancel their subscription after the last episode (most probably now wish they had canceled earlier).
HBO ended up losing about 16% according to this article from complex.com — still bad, but survivable. Now imagine instead of one flagship show, thousands of shows and movies had been taken down in a matter of months.
Before we get all chicken little, let’s break it down a bit. It depends on how far-reaching the definitions of “kids content” are, how aggressive the FTC is with punishing violators and how much support YouTube gives creators. But make no mistake: there is a version of this story that ends with YouTube dying.
What this means for the future of YouTube.
I, like many filthy millennials, love YouTube. It’s a place for content that wouldn’t be considered popular or sane enough for traditional media (I’ve personally fallen in love with this channel where a guy makes knives out of random things like milk or flax seed oil.
From a digital marketing perspective, it serves as a unique ecosystem of diverse interests with an absolutely massive audience: 2 billion active monthly users.
So, consider this conclusion one part prophecy, one part plea and one part personal rationalization.
The privacy of children on the internet is non-negotiable. Collecting data on kids to use for advertising is shady. The answer is not to do away with or ignore COPPA. However, putting the solution entirely on the backs of the creators is also unfair.
The privacy of children on the internet is non-negotiable.
People visit YouTube to watch the content. That content is created by the very people who will be hurt most by these changes. If those creators can’t make a living on YouTube they’ll go somewhere else.
This isn’t 2005 and YouTube isn’t the only game in town. Just as thousands fled Vine after its shut down (RIP Vine) thousands or millions may migrate from YouTube to a new app like Twitch or Mixer. The eyeballs of viewers will follow, along with the ad dollars. I’m sure Amazon and Microsoft respectively would be happy to take them.
Some unsolicited advice for YouTube from a data scientist: if I were in charge at YouTube I’d be focusing less on how to protect the increasingly fragile behavioral targeting ecosystem and focusing much more on contextual targeting. If the value of contextual ads can be increased, then theoretically everyone would be happy. YouTube wouldn’t need to collect user data, COPPA wouldn’t be violated and the content creators could continue to enjoy making a livable wage.
How do they pull that off? I’m not sure, it’s probably not something I could answer as an outsider. What I can say is that the landscape for behavioral targeting is changing and YouTube may have to solve this problem regardless (again, What is Contextual Targeting?).
I hope this turns out to not be a big deal. I hope that in six months I look back at this article and laugh. Just in case, I’m gonna go binge all the exotic-knife-building videos while I still can.
For more expert insights like this, subscribe below or schedule a demo today.
Subscribe to Impressions
Get the latest news, tips, tricks and best practices delivered to your inbox.
Knowing the quality of your digital advertising inventory is paramount to the success of your campaign. Why? High-quality inventory will drive a much higher return on ad spend and ensure you reach your intended audience. Use this short list of questions to determine...
Table of Contents 1. Why It Matters 2. Past & Present vs Future 3. The Problem 4. Confidence Interval 5. The Solution 6. An Example 7. How ToThis is a sequel article to one of our biggest hits of 2019, LumenMath: The Rule of Three, or Should I Keep This Audience...
Table of Contents What It is How To Adapt Add/Combine placements Optimizing ConclusionUpdate 1/30 On Tuesday, 1/28, Facebook rolled out the “Off-Facebook Activity” tool which is an update to what we used to refer to in this article as “Facebook’s Clear History...
Advertising cannabis-based products has always been a sensitive subject in regards to programmatic advertising. In all honesty, there has been little success marketing the cannabis industry as a whole — even using traditional mediums of billboards, newspapers,...
It’s the end of a decade, and for digital advertising … what a decade it’s been. This industry moves so fast that sometimes it’s nice to take a moment to reflect back on where we were 10 years ago. If you are an OG of digital advertising (and let’s be honest, you are...
As you near the end of an advertising campaign, are you scrambling to make sense of mountains of data, striving to deliver a panoramic view for your stakeholders? The problem is not necessarily the quantity of data or the number of data sets, it’s trying to get the...