Recently, there has been a lot of coverage around Facebook selling advertising to Russian company. According to CNN, roughly 3,000 ads were sold to a company in Russia that may have been bought as a way to provide misinformation during the election cycle. The investigation led by Robert Mueller has reached out to Facebook to gain access to the ad data, according to USA Today. But in talking with many folks who are out of the social media marketing industry, it seems that there are plenty of questions when it comes to both the legitimate—and propaganda—use of paid social and dark posts.
Please note, this is not an indictment against a particular political party or candidate, but rather a primer on some of the topics and use cases to help folks better understand the social landscape. I’ve seen plenty of individuals who are social media professionals lack the understanding of these topics. So, if folks in my industry are confused, that people outside of it may have a lot of questions. While the post is long, I wanted to be thorough. I hope that it can be of benefit to journalists, politicians and laypeople as they try to understand how outside forces can use social media to influence our citizens and undermine our democracy.
Evolution to the Paid Social World
The first misconception most people have is that everything that is posted to social media is seen by all the followers. That is unfortunately—especially to all the marketers out there—untrue.
The reason that this misconception exists is because in the early days of social media, it was true that your followers were exposed to all of your message. This is why there was a great push for companies to adopt social media as it was a free communication channel to customers and fans—there was no “ad buy” that you had to make to speak directly to the people. Consequently, many organizations joined social media and the metric du jour was maximizing the number of followers: the more people who followed resulted in more “free” interactions.
Much has changed in the past decade. After getting saturation of both users and companies, social platforms began to become a “pay-to-play” channel in the same vein as television, radio, print and online search. Sure, you could post your message to your followers, but that didn’t mean that they would see that message.
In fact, only a small *fraction* would have ever seen your social post, often ranging from 0.5% to 2.5% of the audience. This is due to the different algorithms—most call it the “Algorithm”—each platform has to surface the most relevant, or engaging, content to their audience. In the same way that you’d have to pay for an ad to be printed in the newspaper or aired on the radio, you now had to pay to make your social update appear in the feed. Paid social was born.
Early on paid social was simply pressing a button to “boost” or “promote” a post. This just meant that more eyeballs (most likely those of your followers) would be on it. But as the platforms have evolved, so has the sophistication in how users are targeted. Facebook has a treasure trove of information, most of it being supplied directly by the user: age, gender, employment, college and family demographics (such as being married or having children). And in the event you follow Facebook pages of musicians, sports teams or actors, the platform has direct data on your personal interests.
But other information is also inferred based on the websites you visit. “What? How does Facebook have my browser history?” you may ask. The answer is that most websites have installed a piece of code (often colloquially called a “pixel”) on their page from Faceboook (e.g. that “Like” button), your session information on that completely separate site is passed back to the social platform giant.
For example, if you’re logged into Facebook and then browse to a home brewing store that has the Facebook pixel installed, all that information will be passed back to Facebook. This enables them to understand more about who you are, your interests and the exact pages you’ve browsed on that site.
All this data and information is then packaged together and advertisers can target people based on certain criteria. Want to target all females in the Bay Area who make $150,000+, don’t have kids and love Justin Bieber? You’ve got 32,000 women to message to. Seriously.
Social media platforms have provided the Holy Grail to marketers to have “personalized” marketing. The “spray-and-pray” mass marketing of old—purchasing billboards, commercials and print ads—has given way to hyper-focused ad campaigns that reach the exact persona that a marketer wants to sell to.
OK, I Understand Paid Social, But What’s a Dark Post?
Now that you have an understanding of how your personal information is being accumulated and used to send you messages, the next topic to explore is that of a “dark post.”
Not everything posted by companies or organizations on social media is accessible to the public. Instead, there is something called a “dark post” or a “sponsored post” that are messages that can only be seen through paid social campaigns. In other words, they are posts that no one would be aware of unless the organization puts a marketing budget behind them. And if there is no money paid to promote the posts—or the campaign budget runs out—no one will be able to see these dark posts.
So, the question you may have is, “Why would anyone spend time creating messages that no one would see?” The answer lies in something called “A/B testing.”
The theory behind A/B testing is to pit one idea versus another and see which one wins. Similar to the scientific method, you’re simply modifying one particular variable and seeing how it compares to the control. If it performs better, you can keep on iterating and iterating until you’ve achieved the best output there is.
Folks have always done A/B testing, though it has always been on a lower scale. Think about back-in-the day when people used phone books to look up phone numbers—savvy business owners would have one number for their yellow page listing and one number for their ad. They could then easily measure how well the ad performed and if it was worth the extra money.
Another way a business could do A/B testing was by placing billboards with varying phone numbers based on their location. The owner could measure the effectiveness of each locale, ultimately informing them of where to focus their ad dollars.
The web—in particular paid social dark posts—has taken this concept of A/B testing and put it to the extreme. While a company may not be willing to flood their public social media presence with tons of variations in ads, they can test all sorts of variations with those dark posts without clogging their organic feed. Companies can make slight tweaks to the copy, headline, link description, feature image and call-to-action button resulting in tens, hundreds, even thousands of dark posts to find the best way to make a consumer take an action.
Instead of putting out a single organic post, a company is rigorously testing and retesting every single variable to see what hits. And when they identify a winner—the one that “converts” in a desired action in the most efficient way possible (read: costs the least amount for a user to take a desired action)—the company goes all in, pushing a majority of their paid social media spend behind that post.
And believe it or not, a single word or a tweaked image can have a monumental effect with respect to online conversion. In a talk I attended at SXSW in 2015, Evany Thomas, a brand writer at Pinterest, said that a simple design/copy alteration was put in place that resulted in an increase of 811% in conversions.
Yes, A/B testing works, and because there are not really any “set up” costs to actually test (i.e. all you are out is time to create the versions rather than having to pay for multiple billboards, print ads, phone numbers as you’d have done in the past), it’s in a company’s best interests to churn out a tremendous amount of variations to see which ones hit pay dirt.
It doesn’t take a lot of effort to create a lot of options and MarTech (marketing technology) solutions can help automate it. The easiest way to get a ton of dark posts is to have a handful of versions for each particular variable of content in a status post. Looking at the below example, you’ll see that there are four main components to a status post on Facebook:
- Status Post: Make the most of your weekend. Make your own beer!
- Featured Image: Guy pouring malt extract into a kettle
- Headline: Craft great beer. Start homebrewing TODAY.
- Link Description: Our all-in-one homebrewing kits are built by brewmasters. So you’re guaranteed great tasting beer, even if it’s your very first brew day! LEARN MORE
Simply coming up with a handful of variations for the status post, featured image, headline and link description will yield many different options to A/B test. In fact, the arithmetic is simple, you take the product of all the variations available to determine how many total ads you can create. In my experience, I’ve typically run three versions of content posts with three versions of headlines and two different images (note: I’ll often keep the same link description). This results in 18 variations (3 x 3 x 2 x 1 = 18) that I can A/B test.
This is a completely reasonable use of having multiple dark posts. Companies are trying to get the best return on their investment, they don’t want to clog their organic feed and they most likely are not doing anything unsavory. In this instance, it’s ok that the dark posts aren’t public—there’s nothing being done to mislead the public and most likely no review needed to be done on what a company is saying to sling some widgets.
Companies with massive budgets could be A/B testing thousands of variations on social media; doing so can give a competitive advantage. In fact, that was one of the keys to Donald Trump’s victory—his digital team was known to A/B test up to 175,000 variations, according to Wired. As Gary Coby, director of advertising at the Republican National Committee stated so succinctly in the article, “The more you’re testing, the more opportunity you have to find the best setup.”
The problem arises when a company—or an outside actor—uses the hidden nature of dark posts and the paid social targeting to deceive.
Pushing Fake News and Propaganda
While I detailed out the proper uses of paid social and dark posts, you’re most likely reading this blog post to understand the nefarious ways that it can be used. The first way is to quickly spread massive untruths. The second way is to keep the darkest propaganda outside of public view while stoking the fire of those who don’t mind being informed by truth.
Using Paid Social to Push a False Narrative
First, let’s examine how many of the fake news utilized paid social to push a false narrative. And to do this, we have to take a trip around the globe to Macedonia. It’s been wildly noted that the little town of Veles was the epicenter of fake news. As an effort to overcome poverty, many teenagers and young citizens turned to churning out a ton of online garbage.
Putting out salacious stories and information resulted in a huge amount of traffic, which in turn was monetized by an excessive amount of digital banner advertisements on their websites. But just as social marketers at Fortune 500 companies know, there is a certain ceiling to organic traffic. If you really want to drive traffic, you have to sometimes pay to get it. And when something is salacious, adding paid media dollars on top of it is like putting gasoline on the fire. It’s a perfect recipe for a post to go viral on social media.
In this instance, a bad actor is painting with a broad brush in spreading fake news and propaganda. They most likely aren’t worried about their message being noticed by watchdog groups or the media, because the lie can be far more wide reaching than the truth. Remember, these folks are primarily concerned with driving clicks to their website to make as much ad revenue as possible.
But it can have profound consequences.
If a lie can travel halfway around the world before the truth puts on its shoes, a lie with a lot of money behind it has done half-a-dozen laps. The damage can be done so quickly and debunking a falsehood that has spread so far-and-wide is more than a challenge. Additionally, the very act of debunking the lie may only solidify it as truth.
Let’s look at this example: “Joe claims, with no evidence, that Sandy embezzled over $50,000 from the non-profit organization.” Even though you are explicitly trying to paint Joe as a liar, the very message of “Sandy embezzled over $50,000 from the non-profit organization” is repeated, helping to perpetuate that Sandy may indeed be a thief.
Our next example is not concerned as much in ad revenue and more in truly shaping the course of a sovereign nation. This threat is probably more concerned with keeping their message hidden, something they can do with dark posts and hyper targeting.
Obfuscating the Message from Watchdogs with Dark Posts
In the past, there was always public visibility—and a record—of what advertising was being purchased in print, radio or television. In the digital world, there are similar records of social media posts that are made organically by companies or individuals. However, there is really no easy way to track down what has been said via dark posts.
This presents an incredible ability to deceive and spread downright lies to a very targeted group of people without ever being seen by a watchdog, consumer advocacy group or the media. In fact, you could theoretically guarantee it by either targeting with *very* specific parameters or by somehow accessing voter data and targeting specific email addresses. Doing so would limit the exposure and ensure your message is very subversive.
Furthermore, outside threats could easily A/B test multiple messages on that group without fear of being seen. The data received on the different variations can further inform their content strategy, seemingly giving the bad actor tons of information of what resonates with a particular audience.
The goal of those entities using dark posts in very targeted, surreptitious ways is much different than those who are trying to generate ad revenue. Rather than trying to make a profit, they are looking to destabilize our very democracy by creating doubt, animosity and paranoia among our citizens. One very current example was featured on CNN where Russian actors purchased ads about Black Lives Matter to “sow discord” in the communities of Ferguson and Baltimore.
It is not a far stretch to see how foreign nations and bad actors could use a paid social media campaigns to spread fake news and propaganda to specific swing counties and demographics that could ultimately be very influential in a close election or even create chaos to weaken our institution of democracy.
Based on the activity in 2016, the one sure bet for 2020 is that fake news isn’t going anywhere. While Facebook is flagging fake news posts and now banning those pages from purchasing ads through their platform, we are at the beginning of a cat-and-mouse game. Plainly put, there is too much to be gained by both the small-time players and adverse nations. I hope that this information was helpful for those reporting on how social media may have impacted the 2016 election and those simply seeking information. If you have any questions or need further clarification of the terms, please feel free to leave a comment.