Is Facebook Too-Big-To-Fail?
The Fake News Issue Demands We Ask That Question
Facebook allows fake news to spread on its platform. Potentially it influenced the U.S. election’s outcome in doing so. That raises questions: What role does Facebook play in the news ecosystem? Is it systemically relevant? Too-big-to-fail even? An analytical look at the network’s influence on users, media and the public sphere.
I think some part of the fake news debate surrounding Facebook is overblown. Insofar I stick to my earlier remark that claims like Facebook made Trump are certainly exaggerated and, worse, try to find an easy scapegoat instead of tackling the real issues.
That said, I kept following the reporting on the issue and changed my initial position. While I still think it’s dangerous to blame a single actor (Facebook) for the outcome of this complex election, by now I admit it might have had a larger influence on it than I initially acknowledged.
Apparently fake news, at least the popular bits, were heavily skewed towards Pro-Trump/Anti-Clinton pieces. In the three months leading to the election it outperformed posts by serious media in terms of engagement on Facebook. A particularly nice take by Max Read is his framing that Facebook has turned into an instrument that allowed DDoS attacks on civil society by being the grounds on which right-wing propagandists and Macedonian scammers likewise flooded the public sphere with junk news.
While I still think the questions I posed here ought to be answered before agreeing on a final verdict, Rick Webb’s argument that even small percentages matter is pretty plausible to me (and his piece well worth a read!).
But I don’t want to write about the percentages to which Facebook assisted in Trump’s election. Because regardless of the answer, the overall development’s direction is worrisome enough. The influence Facebook has as the intermediary between people and the media is huge without a doubt. Take the relatively recent fake news issue and add to it the broader — but long-term probably more important — filter bubble issue (nicely summarized by Tobias Rose-Stockwell), and what we get is a mixture with very frightening potential. Brian Phillips, in his brilliant essay, is on point:
The sensational headline that’s been thrown at Facebook’s misinformation problem asks whether fake news cost Hillary Clinton the presidency. But you don’t have to believe Facebook got Trump elected to be a little chilled by its current estrangement from fact. One of the conditions of democratic resistance is having an accurate picture of what to resist. Confusion is an authoritarian tool; life under a strongman means not simply being lied to but being beset by contradiction and uncertainty until the line between truth and falsehood blurs and a kind of exhaustion settles over questions of fact. Politically speaking, precision is freedom.
…
Authoritarianism wants to convince its supporters that nothing is true, that the whole machinery of truth is an intolerable imposition on their psyches, and thus that they might as well give free rein to their fantasies.
Now, I wouldn’t accuse Facebook of being authoritarian (given their willful inaction). To borrow the words of James Allworth and Ben Thompson, we just witnessed sins of omission rather than commission. But at scale the effect of either is just as disastrous.
Which leads me to the questions I’ve been grappling with the last few days: Is Facebook too-big-to-fail? And if so, which reasonable ways forward exist?
I don’t intend to give a conclusive answer. Rather, I want to present the key facts on which a discussion can unfold and share some thoughts on the second question.
But first, let’s make sure we are on the same page terminology-wise. Too-big-to-fail is a concept that rose to popularity during the ‘07/’08 financial crisis. Back then, some banks were basically broke but eventually bailed-out by governments because their bankruptcy would have endangered the entire global financial system. They were deemed ‘too-big-to-fail’. On a more abstract level, thus, the term refers to single entities within a system which play such an integral role in the system’s survival that their breakdown would likely cause the entire system to follow suit. The existence of such too-big-to-fail entities is the result of a poor risk distribution strategy.
So, with that in mind, let’s look at Facebook and the role it plays within the system of news, the public sphere and opinion forming.
Facebook’s scale
In case you didn’t check your Facebook numbers in a while:
- 1.79 billion monthly active users
- 1.18 billion daily active users
- 1.09 daily active mobile users
(all as of September 2016 from Facebook)
Also, 1 billion users accessed it exclusively via mobile.
Geographical distribution might be worth taking a look at, too:
(both via Facebook Q3 2016 results)
So — not surprisingly — Facebook is quite big. Certainly an indicator when it comes to answering our question. But certainly not enough. What we really ought to understand is the influence Facebook has on media consumption.
Facebook’s systemic relevance
Being too-big-to-fail is all about systemic relevance. We’ll first take a look at the demand side (users) and then the supply side (the interplay between Facebook, news media and advertisers).
Reliance on Facebook as a news source
Let’s start with some numbers. We have some pretty decent, U.S.-specific numbers from the esteemed Pew Research Center:
- 44% of American adults get news from Facebook
- 64% of U.S. users who get news from social media sites only get them on one site (to be clear, this doesn’t map 1:1 on Facebook users)
An interesting question is what percentage of Facebook users only relies on Facebook for news. The following graphic is the closest to an answer but alas it’s fuzzy.
Apparently, the news source most of Facebook’s users also get news from is TV (39%). But since we don’t have the raw data we don’t know the overlap between the numbers. So we are not that much smarter. But we’ll come back to the question in a moment.
What stands out to me so far: The dictum ‘people live inside Facebook’ is apparently true, at least when it comes to online news. Only 33% of Facebook users in the U.S. claim to access other news websites or apps.
Another source I looked into is the Digital News Report annually published by the Reuters Institute. The survey was conducted in 26 countries with at least 2.000 participants in each. Decent data that allows us to look at the international picture. Some findings from the 2016 edition (PDF report):
Internationally, Facebook is the top social network for news; 44% of survey participants state they used it as a news source within the last week. There are some notable differences between countries though. Some examples:
Brazil: 69%
Greece: 68%
Canada: 46%
US: 45% (pretty much in line with the Pew number; good sign wrt reliability)
UK: 28%
Germany: 27%
Japan: 16%
(you can look up all countries in the interactive tool here)
Notably, Facebook is the most used social media news source in all surveyed countries except Japan. Unsurprisingly, young people rely more heavily on Facebook as a news source: 55% of 18–24 year-olds across all surveyed countries.
The Reuters survey also asked people what their main source of news was. In the U.S. social media (note: the data isn’t Facebook-specific!) are the main news source for 14% or respondents. The EU average is 10%, though with significant differences between countries. In Portugal, the high mark, 16% get news mainly from social media compared to 6% in Germany. The most social-centric news consumers live in Brasil and Australia at 18%. The number increased year-over-year in almost all countries (Finland is the only exception).
In line with the findings for general Facebook usage, young people rely more heavily on social media as their main news source: 28% of 18–24 year-olds (all 26 countries).
As always, numbers are up for interpretation. What appears clear is that only a minority of people relies primarily on social media — and in extension Facebook — as a news source. However, it’s not an insignificant minority as of today and the numbers are increasing, both because of changing habits (as year-over-year growth suggests) and demographics.
The Online News System
But there is more to our question than the numbers. Discussing systemic relevance also requires us to understand the system at hand. In our case this means looking at the interplay between users, news organizations, advertisers and Facebook/platforms.
The key issues at play there are:
- limited attention
- user experience
- the power of aggregation
- user lock-in
- scale & monetization
People have limited time and, thus, attention. Therefore, they tend to spend their time at the places which offer the best user experience. Platforms which instead of creating content use aggregation — the bundling of news content created by others— can focus on the user experience. They gain an important advantage: users’ time. Because attention is at the core of the competition for advertising dollars, platforms take active measures to lock people into their system. If they succeed, they scale. And size attracts advertisers’ money.
That’s essentially Facebook’s playbook. In his piece on the fake news issue, Ben Thompson aptly summarizes the result:
In Aggregation Theory I described the process by which the demise of distribution-based economic power has resulted in the rise of powerful intermediaries that own the customer experience and commoditize their suppliers. In the case of Facebook, the social network started with the foundation of pre-existing offline networks that were moved online. Given that humans are inherently social, users started prioritizing time on Facebook over time spent reading, say, the newspaper (or any of the effectively infinite set of alternatives for attention).
It followed, then, that it was in the interest of media companies, businesses, and basically anyone else who wanted to get the attention of users, to be on Facebook as well.
It’s hardly an exaggeration to say that almost every media company today relies on Facebook to reach its audience. And Facebook isn’t shy to leverage that power in several ways. Most notably their Instant Articles feature encourages publishers to put their content directly on the site. So, instead of driving traffic to the media’s websites users no longer need to leave the immersive Facebook experience. It’s working to such an extend that a significant amount of Facebook users in several developing countries isn’t even aware they are using the internet. To them, it’s all Facebook.
What I noted elsewhere with regards to Google equally applies to Facebook: The ideal state from their perspective is to become the single touchpoint people use. To be sure: The numbers show that’s not currently the case. However, it is clearly the direction of the news system’s development. Moreover, a self-reinforcing feedback loop might be in play.
Following the money
We must not underestimate the relevance of the news system’s economics. The most important revenue source for news media is advertising. Advertisers go where the people are. Thus, they are on Facebook (and Google; while these two are growing there are indicators that the rest of the online advertising industry is shrinking).
This alone puts publishers at a disadvantage. Add on top the superior user data Facebook owns and it’s basically pointless for any single publisher to try competing. There is a clear incentive for them to instead use the tools Facebook offers and share in the advertising dollars. In order to do so, they of course have to play by Facebook’s rules and feed them their content. That’s the structure of a virtuous cycle (or a vicious one, depending on who you are).
Alas, I didn’t find any data that shows the current revenue composition of digital publishers (even Pew has no specifics). But it wouldn’t surprise me if it increasingly tilted towards platforms and Facebook in particular. Especially because advertisers now increasingly spend on mobile and Facebook exceeds there.
(source: Pew Research Center)
The end is nigh
Let’s sum it all up. On the demand side, users are neither fully nor solely relying on Facebook for news consumption. Things are more complicated on the supply side though. The dynamic between news media, Facebook and advertisers suggests that Facebook’s key assets — attention and the following ad dollars — give it a considerable leverage. This results in Facebook having substantial but not quantifiable power in the news system. Based on current trends the situation is likely to develop further to Facebook’s advantage.
So, is Facebook too-big-to-fail?
As I stated earlier I won’t give a final verdict (but I invite you to share your conclusion in the comments!). I think you can argue both cases. Right now, I think it’s a close call either way. When projecting into the future, though, I believe there is legitimate reason to worry. If the trendline continuous its current course (which is no given) Facebook’s influence on users and the media will reach a critical level that certainly deserves being called systemically relevant.
Hence, let me briefly address which responses this should and shouldn’t provoke.
The obvious call would be to ask for regulation. But active government intervention is a dangerous path when it comes to news. It serves an important function for the public sphere: being the basis for discourse and opinion forming. To quote Thompson again:
it is a frighteningly small step from influencing the world to controlling the world
…
If users won’t seek out the “right” news sources, well, then someone ought to make them see them. It all sounds great — and, without question, a far more convenient solution to winning elections than actually making the changes necessary to do so — until you remember that that someone you just entrusted with such awesome power could disagree with you, and that the very notion of controlling what people read is the hallmark of totalitarianism.
His proposed solution is certainly a step in the right direction: more transparency. Having more insight into the workings of Facebook’s algorithm as well as decision making processes would certainly improve things. But I’d go one step further.
When an entity is too-big-to-fail the natural solution is to scale it down. In the case of companies this would usually mean breaking it into pieces. But since what’s concerning here is less the company but the platform Facebook, that’s not really an option. After all, the undeniable value it provides is closely related to its scale. So breaking it up is no option.
Instead, I think about power distribution. The scary thing we are facing is the incredible amount of power that comes from potentially controlling what people read, watch, see and listen to. All this power in the hand of one entity — a for-profit company at that — is certainly nothing to wish for. The downside is suddenly apparent when considering today’s news that Facebook developed a censorship tool in order to appeal to the Chinese government. So the question I’d pose is: How can we distribute that power?
I don’t have a fully thought-through strategy but some initial ideas in varying degrees of out-there-ness (all of which Facebook would obviously hate. So getting there would almost certainly demand some intervention; but I regard them as smarter than the approaches I tackled above):
- Giving users full control over their feed and the decisions it makes about what to see and what not to.
- Allowing third-parties largely extended API excess so they can develop new interfaces (ideally with better customer experiences, e.g. fine-tuned to local preferences) on top of Facebook and it’s data. Think of apps for news discovery, completely different interfaces, alternative feeds, advertising solutions etc. This would basically render Facebook a tool for identity management and a content database that enables all kinds of competing applications.
- Building a governance structure that gives users a say in management decisions
- What else?
