Is Facebook Filtering Content for More Ad Revenue? Guest Post from Craig Wilcox of Simplur

Is Facebook Filtering Content for More Ad Revenue?

The treatment of recent major stories by Facebook highlights how our country’s dominant social network can amplify likable stories and filter out uncomfortable stories and challenging ideas. Facebook is optimizing user experiences away from important social information that might otherwise lead to substantial social progress. The question that naturally arises is whether Facebook has any duty to show us more critical information. Here I’ll demonstrate how Facebook likely does this, spin out an answer to whether they have any moral duty to change it, and propose a better site and business model for making social change happen.

Ice Buckets and Icebergs

Over the summer of 2014, the prevalence of the ALS Ice Bucket Challenge became almost inescapable on Facebook. By mid-August, over 1.2 million challenge-related videos had been shared on Facebook. Total fundraising eventually went over $100 million.

Also trending during this time was the August 9, 2014 police killing of teenager Michael Brown in Ferguson, Missouri. On Twitter, which is unfiltered, the hashtag #ferguson dominated. On Facebook, it was virtually invisible.

The proverbial Titanic versus iceberg comparison is good here. While Facebook was dumping ice over its head in celebration, we were storming toward the iceberg of police brutality. Other events during the summer showed we had run straight into it, causing fatal sinking events throughout many urban areas.

Protests continue in Ferguson and other areas (#moralmonday on 10/13/14, for instance), even though many who get their news on Facebook may be completely oblivious to them. Americans users spend an average of over 40 minutes per day on the site, and according to Pew Research, at least 30% of users depend on the site to gather news. Since those same users spend only 1.1 minutes per day on actual news sites, it’s easy to argue Facebook filters actually exacerbate ignorance among its users. Not only do they not see the stories on Facebook. They likely also don’t know this and therefore don’t pursue other avenues for news stories.

EdgeRank, Affinity, and Filtering

Getting back to Ferguson, we know the story would have been downgraded by Facebook’s EdgeRank algorithm. Facebook developed the EdgeRank algorithm precisely to determine whether a post should be shown to the user’s network or not. The major factors of the algorithm are described here well, and they constantly change, but we will focus particularly on the “affinity” signals.

Affinity is the most important EdgeRank signal for keeping users consuming ads on the site. Google gives us this definition for affinity: “a spontaneous or natural liking or sympathy for someone or something.” In essence, Facebook keeps us on the site, and keeps us coming back, because it’s a short-term anti-depressant. Its algorithms are designed to show us things we like because things we like make us feel good – temporarily, at least. (Using Facebook for too long seems to have a depressive effect, as people realize they are seeing others doing happy things while they’ve been on Facebook for the past hour.) Feeling good on Facebook makes us spend more time on it.

Facebook is, quite simply, internet crack. Increasingly, users are getting a “hit” of Facebook through mobile devices many times each day. In 2013, the average smartphone user checked Facebook 14 times per day. (http://mashable.com/2013/03/27/facebook-usage-survey/) The affinity algorithm would have calculated ALS Ice Bucket Challenge posts to be the purest form of the drug. Though it is certainly more sophisticated in its methodology, the algorithm would have seen both a high percentage of people liking each post and a high number of people sharing each post. It kept showing more and more of these videos as we kept clicking the like and share buttons.

Ferguson, on the other hand, was far less “likable”. Many have called for a prominent “dislike” button for stories that affect us negatively. However, disliking something leads to less time on site, as it would break the most basic part of the affinity model: likeness. If disliking something meant more time on site, Facebook would certainly have it by now.

Instead of liking a Ferguson post, then, users would be more likely just to share it. But even a non-genius mathematician would be able to figure out that something with a low like percentage and a high share percentage should be filtered out for most users.

If A Tree Falls on Twitter…

As a result of sharing on unfiltered networks like Twitter, we will still likely see action to prevent more shootings and better police behavior. Social action usually arises from problems being highlighted, and our country has shown a remarkable ability to mobilize for solutions. The Ferguson police department will clearly have to make some changes, for instance. Another issue that has been raised as a result of protests is whether programs that militarize our police force are harmful. Since the shooting happened behind the squad car’s dashcam, some communities will now be requiring bodycams for police in addition to dashcams. Perhaps most importantly, the culture of training our police force is likely to be reformed.

However, if Ferguson would have been shown as much on Facebook as on Twitter, wouldn’t those calls for action have been far stronger? Facebook remains the dominant force in social networking, even though Twitter has made enormous strides. Doesn’t Facebook understand it has a responsibility to show us important information? As Eli Pariser pointed out in his TED talk on the “filter bubble”, Facebook CEO Mark Zuckerberg has said, “A squirrel dying in front of your house may be more relevant to your interests right now than people dying in Africa.” So, apparently not.

Even though I find this view reprehensible, I don’t believe Facebook has an obligation to show more news stories by altering its affinity algorithms. (I do believe it should alert people to filtering better, just like it has done or been forced to do with privacy settings.) The business model it chose and the responsibility it thus has to its shareholders is incompatible with showing more socially impactful information. In other words, it wasn’t built to make the world better. It was built to make money by connecting people to things they like more quickly than anything else we’ve seen on earth. Facebook’s algorithms work perfectly well for what it was built to do.

Facebook is protecting the user experience by filtering this information. Most people see around 20% of the posts available to them. For posts like those about Ferguson, the reach is likely much lower than 20%. I would predict that my social action posts on Facebook – posted via Simplur – are shown to no more than 2% of my network – the same 8 or 9 people each time, it seems. Even though my ego takes a blow from that, realize Facebook just doesn’t have room for it. For people who are interested in making a social impact, it’s time for a new platform.

Simplur: A Social Network for Social Impact

A number of years ago, this revelation led me to imagine and start building Simplur – a social network that is explicitly built with social and political action in mind. People can use it the same way as Facebook or Twitter. The big difference is with the business model, which prompts us not to filter social action, but instead to encourage it.

Simplur’s business model is based on showing users ways to get involved with political and charitable causes that match the social action they are passionate about. We take a small percentage of any posted transactions, and this should be a sufficient revenue source to power our efforts.

In the background to the social networking experience, Simplur has built online giving tools that show up in the newsfeed, and we have created a feature called Pluralized Funds that allows users to gather money together to fight against what is generally called “big money” in politics. It’s a tool for the 99% to be able to give money the same way a PAC or corporation would and thus have more influence than they would by giving separately.

On Simplur, between cat videos, the ALS Ice Bucket Challenge, pictures of kids and well-plated food, our users see ways to prevent more Fergusons as shared by other users. We have enabled giving for over 1.6 million causes, and we actively encourage users to donate and meet others who are interested in the same causes.

During its last quarter, Facebook reported it had advertising revenues over $20 per U.S. and Canadian user on an annualized basis. Simplur will not make it to those lofty levels, but becoming billionaires is less important to us than making a difference while lowering fundraising costs for non-profits. As Craigslist has shown, it’s entirely possible to run a large internet site without maximizing advertising revenues.

And, of course, we will still actively encourage viral movements like the ALS Ice Bucket Challenge. On Simplur, the idea is not Ferguson or ALS Ice Bucket Challenge – we want both. We believe a social network can and should be both fun and socially conscious. Those are not mutually exclusive goals.

Unless the rare liked and shared (“viral”) campaign like the ALS Ice Bucket Challenge can be created, political and charitable causes will continue to have a difficult time making Facebook work for their interests. While Facebook remains an important place to have a presence and to learn about the zeitgeist of social networks, it may very well be considered closed for business for those who want to use it for social change.

Simplur is not Facebook, but we add more and more features with each day. For those who would like to steer our ship away from the icebergs, it’s worth trying. Instead of investing our revenues in algorithms that filter away impactful opportunities, we’ll be making sure it’s seen and that people have a chance to gather up for change.