Support This Blog On Patreon!

Support this Blog!

All I ask is a $1 a month. (But more is great too.) If you find this content to be beneficial, interesting or just a fascinating peek into true insanity please donate.

Saturday, November 11, 2017

Is Facebook More Like a Newspaper or a Video Game?

If you prefer to listen rather than read, this blog is available as a podcast here. Or if you want to listen to just this post:

Or download the MP3

Over the weekend I listened to A Connecticut Yankee in King Arthur’s Court (as narrated by Nick Offerman.) It’s an enjoyable book, though something of a polemic, and easier to understand if you know that Twain hated Sir Walter Scott (among other things he blamed him for the Civil War). As you may or may not know, Scott was a well-known novelist of the time who romanticized the entire medieval period, and, when you read Twain’s novel, it’s apparent that it was born out Twain’s dislike of Scott specifically and of the idealization of the medieval period more generally.

One of Twain’s major goals was to show how backward everything was during the medieval period, and how awful things were for the great majority of people. Consequently one of the major themes of the book is the importance and wonder of progress, and on that front I may revisit it in more depth, but for now I just want to pluck one fact out of the book to set the stage for this post’s subject.

If you’re unfamiliar with the story, an engineer is sent backward through time to the era of Arthur, the Knights of the Round Table, and Camelot. Where, in the course of various adventures, the engineer attempts to modernize ancient Britain by implementing such things as the telegraph (and telephone), a new monetary system, and the abolition of slavery. All of it spiced up with the liberal application of dynamite. But of all the things this engineer considers important, Twain lays particular emphasis on the creation of a newspaper. As I recall it’s one of the first, if not the first thing the engineer turns his mind too once he has a free hand.

The importance of newspapers specifically and the free press more generally, was not only important to Twain, of course, it has been a major feature of American ideology going all the way back to the founding. And it has generally been seen as something which by itself counterbalances all manner of possible abuses. For example, this quote from Jefferson sums up the role of “papers” nicely:

The people are the only censors of their governors: and even their errors will tend to keep these to the true principles of their institution. To punish these errors too severely would be to suppress the only safeguard of the public liberty. The way to prevent these irregular interpositions of the people is to give them full information of their affairs thro’ the channel of the public papers, & to contrive that those papers should penetrate the whole mass of the people. The basis of our governments being the opinion of the people, the very first object should be to keep that right; and were it left to me to decide whether we should have a government without newspapers or newspapers without a government, I should not hesitate a moment to prefer the latter. But I should mean that every man should receive those papers & be capable of reading them.

This was written while Jefferson was ambassador to France, and if you’re familiar with his later long running battle with Hamilton (which occurred mostly in the newspapers of the time) it’s possible he may have eventually moderated this absolute support. But I’ll have more to say about that later.

To our examples of Twain and Jefferson we could add the First Amendment, of course, and also dozens of other historical quotes all supporting the freedom of the press. Though, as I said in the past, despite all this historic support, there have been some people who have recently started to question unlimited freedom of the press. Still this mostly comes up with reference to hate speech. When you’re talking about being informed about politics, almost everyone from Jefferson, down to the present day has felt that absolute freedom to discuss politics is central to the American ideology and particularly central to the workings of democracy. At least... everyone thought this... until Russia came along... and started buying ads on Facebook…

Okay, I might be exaggerating the impact, but the alarm over the issue is interesting. Particularly the question of where Facebook fits, when we’re talking about freedom of the press and newspapers and all the things which have been so important back to the very beginning of the republic.

When speaking of where social media fits, if nothing else, it’s definitely clear that the rules of the game have been dramatically changed. To illustrate this I’d like to start with looking at the money spent by the Russians. Lots of people go on and on about the Russians tampering with the election, but how much money and influence were they really throwing around? For me this appears to be the part of the story getting the least critical attention, but the part which is potentially the most fascinating. To keep things simple let’s look just at what was spent on Facebook.

From what I can tell there are two numbers floating around, $100,000 and $46,000. I think one number is earlier in the year, and one number is right on the eve of the election, but I’m happy to add the two of them together to get a grand total of $146k being spent by the Russians on Facebook. In fact let’s be even more conservative and assume that some spending hasn’t yet been uncovered and double it, and then round the whole thing up to $300k.

Having come up with a total, the first question I want to ask is, how does that compare to what the candidates themselves spent on Facebook? Well the number there, as far as I can tell, is $81 million for both candidates, which means the two candidates outspent the Russians (even using our very conservative figure) by 270x (or 3/10ths of a percent.) And, as I said this is a conservative estimate, TechCrunch, a site, which as far as I know, doesn’t have any conservative leanings, looks at the more narrow pre-election spending and concludes the Russians were outspent by 1,760x (or 6/100ths of a percent.)

But wait? You may be saying. I heard that the Russians reached 126 million people, isn’t that 40% of the country? Surely that has to represent more money than the $146k or even the $300k we’re talking about? Perhaps. As far as I can tell Facebook will sell you impressions (a fancy way of counting every time your ad gets loaded whether or not the person even notices it) for a half a cent. At that rate 126 million impressions would have cost them $630k, so still very much within the realm of numbers I’m talking about and it only takes our worst case scenario from 270x, to 129x (still less than a percent) as much spending. But also, Facebook gives you a break if your content is particularly viral, meaning that maybe they got the impressions for less than that. Also, I assume that Facebook, at the point where they have to appear before congress and explain themselves, would make sure they reconciled the amount spent with the number of people reached, but maybe not. Either way I don’t think it changes the fact that the Russians spent a relatively tiny amount. This comparison becomes more extreme when you look beyond the candidate’s Facebook spending to the total campaign spending, which clocks in well north of a billion dollars. (This site gives the total money raised by both candidates at $2.35 billion.) Of course then you’d want to look at total spending by the Russians across all platforms, but you still get a situation where the money spent by the candidates is hundreds of times greater than that spent by the Russians.

All of this leaves us with two possible conclusions: Either, the Russian money was a drop in the bucket, and it had no effect on the outcome of the election. (And everyone should get over things.) Or, social media spending, particularly of the kind the Russians did, is disproportionately effective, and that for a measly 150k (or 300k, or 600k) they were able to buy the presidential election for Trump, when it otherwise would have gone to Clinton.  Now, to be fair, this was a close election, and in close elections you have the benefit of being able to point to any single factor and credibly claim that it could have swung the election (2000 is great for that sort of thing). And thus, I suppose, it’s even possible that there’s a third option. That each dollar the Russian groups spent on Facebook was about as effective as a dollar spent on TV ads, or canvassing, or what have you, but that the race was close enough that it still made the difference in who won.

If it’s the first, that the Russian spending made no difference, then there’s not much of a story, just the typical Monday morning quarterbacking that’s going to happen after any close election. Democrats can’t accept that they lost “fairly” and so they focus on nefarious outside influences to explain the election. It wasn’t that Clinton should have campaigned more in Michigan (or at a minimum kept her TV ads there running in October), it was the evil Russians (cue Boris and Natasha.)

This is certainly a possible explanation, and perhaps even the most likely, but it’s not very interesting, so for the rest of the post I’m going to assume that social media did have a disproportionate impact on the race, and to be fair there is more evidence than just the Russian angle. When even the Economist is running a cover story titled Social media’s threat to democracy you have to figure that I’m not the only one who thinks that social media spending might have been disproportionately effective. For example, if we look beyond the Russian angle, many people think that it was Trump’s (technically Jared Kushner’s) mastery of Facebook that explains why he won. This section from The Economist is particularly interesting:

The Leave campaign...experiment[ed] with different versions [of Facebook ads]... dropping ineffective ones. The Trump campaign in 2016 did much the same, but on a much larger scale: on an average day it fed Facebook between 50,000 and 60,000 different versions of its advertisements...Some were aimed at just a few dozen voters in a particular district.

As I said in the beginning, there’s a strong bias in American towards considering the press, and particularly the newspapers, to be the good guys and going out of our way to give them as wide a latitude as possible, particularly in reporting political matters. If that’s correct and the press are the good guys, what changed with Facebook? Why is it different than a newspaper? Why is its effect malevolent where previously, on the balance, the press was considered to have a benevolent effect? (There is an argument that it doesn’t, but for the purposes of this post we’re going to assume that Twain and Jefferson were correct.)

I think that quote about the 50,000 ads a day gives us a pretty good idea of the difference.

The first place where most people are inclined to place blame is the hyper-partisan character of the ads Trump and the Russians (and to a lesser extent the Democrats) were showing on Facebook. And I agree this is tempting target, but remember how I said that we would return to a discussion of Jefferson? Well, when Jefferson, who, remember, was a strong supporter of freedom of the press, was waging his titanic battle against Hamilton. That battle mostly took place in the newspapers of the time, and the partisanship and venom of those newspapers makes our own disagreements seem pretty mild. And these weren’t obscure newspapers, these were the major papers of the time. Thus I don’t think hyper-partisanship is a very good candidate for the difference between newspapers at the founding of the country and social media today.

The next place people look, and a place I’ve gone to myself, is the idea that Facebook is just too big. Senator Al Franken (yeah the guy who used to be on SNL) just recently gave a speech about this issue. Wired titled it the Speech that Big Tech has Been Dreading, and in it Franken not only called out Facebook, but also Google and Amazon, saying the following:

Everyone is rightfully focused on Russian manipulation of social media, but as lawmakers it is incumbent on us to ask the broader questions: How did big tech come to control so many aspects of our lives?

Now whether Facebook has too much control or whether they’re big enough to be a true monopoly is certainly up for debate, but you can see where a large number of people get a significant amount of their information from this one source. In other words, there are a lot of people who spend 4 hours a day on Facebook, but don’t spend five minutes reading something like New York Times. Now if Facebook ended up showing these people a broad selection of news from all over the political spectrum this dominance might not be a problem, but as we all know, Facebook targets you with ads they know you will like. And beyond that they provided Trump and the Russians with the ability to target ads down to the level of the individual. As the Economist pointed out, in addition to those 50,000 ads, they were also running ads which were aimed at just a few dozen voters in a particular district. All of which brings me to my next point.

When you read the front page of the New York Times, or even of the local paper you know you’re reading exactly the same front page as everyone else. This has largely carried over to the web. When I go to I see the same thing my wife sees when she goes to But this is not what happens at all with Facebook, as the example of targeting a few dozen individuals indicates. But people don’t realize this. People see something on Facebook and they naturally assume that this is more or less what everyone is seeing. Consequently they are far less likely to question whether it’s true or not. And even if they do question it, or ignore it, or if it’s ineffective in any way, then the Trump team still has 49,999 tries, each DAY, to get it right.  All of which is to say that Facebook (and most social media) is targeted and responsive in a way that makes it a completely different animal from traditional media. The Economist describes it thusly:

The algorithms that Facebook, YouTube and others use to maximise “engagement” ensure users are more likely to see information that they are liable to interact with. This tends to lead them into clusters of like-minded people sharing like-minded things, and can turn moderate views into more extreme ones. “It’s like you start a vegetarian and end up a vegan,” says Zeynep Tufecki of the University of North Carolina at Chapel Hill, describing her experience following the recommendations on YouTube. These effects are part of what is increasing America’s political polarization, she argues.

Even taking into account this increase in polarization, if that’s all there was to it, we’d probably be okay. But on top of everything else, not only do Facebook and similar sites, disseminate hyper-partisan information, act as a near monopoly, convince isolated people that they’re part of a community, and increase polarization, they are also are doing their very best to make all of this as addictive as possible.

This is usually the point where someone comes along and accuses me of being just another old guy who thinks the world is going to hell in a handbasket and just wants the damn kids to get off his lawn. And maybe that is exactly what’s going on. Here’s what the article from The Economist had to say about this possibility:

Social media are hardly the first communication revolution to first threaten, then rewire the body politic. The printing press did it (see our essay on Luther). So did television and radio, allowing conformity to be imposed in authoritarian countries at the same time as, in more open ones, promoting the norms of discourse which enabled the first mass democracies.

Several things are worth pointing out from that quote. First, there at the end, we have another example of past media serving to improve democracy. Second if you boil down the question to whether social media represents a communication revolution, and separate it from the effect it may or may not have had in the most recent election, I think most people would not hesitate to declare it revolution. If that’s the case, perhaps we should move away from considering the narrow question of what the Russians did or didn’t do, since we may be too close to the issue, and examine other examples of communication revolution. The quote mentions two, the more recent TV and radio revolution and the revolution in printing at the time of Martin Luther.

For those who aren’t up on their history. The reason they had an essay on Luther is that we just hit the 500th anniversary of the nailing of the 95 Theses to the Wittenberg church door (if that in fact happened). And as they allude to, one of the big reasons the Protestant reformation happened when it did was the invention of the printing press. Luther himself was a master of the medium and in the 1520s, he was responsible for more than a fifth of the empire’s entire output of pamphlets. At the time one churchman said, “Every day it rains Luther books. Nothing else sells.”

As you also may or may not recall the Protestant Reformation resulted in one of the bloodier periods of European history (the 30 Years War, the French Wars of Religion, the English Civil War, etc.) So that’s one of the revolutions. On the other hand, the TV and Radio revolution was almost entirely peaceful and we can always hope that even if social media does represent a revolution it will be more similar to the revolution brought on by TV and Radio than the one brought on by the printing press. Though, I fear that when you look at ease of control, social media is a lot closer to the decentralized revolution of the printing press, then the centralization of TV and radio.

The only question left, assuming you agree with me thus far, is what we should do about it? The Economist asks the same question and comes up with this response:

What is to be done? People will adapt, as they always do. A survey this week found that only 37% of Americans trust what they get from social media, half the share that trust printed newspapers and magazines. Yet in the time it takes to adapt, bad governments with bad politics could do a lot of harm.

I agree. People will adapt. But remember that Facebook and the other sites can also adapt. The Russians can adapt. And we already read about the 50,000 adaptations the Trump campaign was making every day. I have no doubt we will eventually adapt, but in the meantime we’re trying to hit a moving target, and a lot can happen while we’re working it out.

Finally, given the evident success of this tactic, how much more money and how much bigger is this problem going to be in 2020?

If this post made you hate social media, or if you already hated social media, consider donating to something that isn’t, this blog. Sure I’m not as reliable as the newspaper, but I’m better than your crazy uncle.


  1. A few other things to think about:

    You have massive armies of 'tweet bots', essentially fake accounts that shared, retweeted, liked, ads and memes over and over. The actual paid ads then are leveraged multiple times over by these accounts. I'm not sure about Twitter's terms of service but Facebook's 'rules' do not allow fake accounts, they insist that users register with their real names. I've never worked with Facebook ads but I have a little bit with Google. Google's ad rates drop if your ad gets more clicks than other ads, that's part of their philosophy of making all their results as 'relevant' as possible for users.
    We have at this point well documented collusion of the Trump campaign with Russia. We have high and medium level officials in the campaign who knew Democratic emails were hacked before it became public, we have requests from the campaign to Russia to hack even more emails. Even if you dismiss Trump's public request for Russia to hack emails as the usual delusional rantings of a somewhat senile fool, the two go together in a way that a simple ad buy cannot.

    Why have never seen KGB, MI6 or China running political ads on TV during elections Because it is illegal? Maybe we should just go global and let anyone who wants to buy ads or contribute to our campaigns do so. Or maybe we shouldn't. Russia's interference wouldn't have been as disruptive if, say, Canada, France, Mexico etc. all had their say.

    Russia is no longer the powerful country but it does have an agenda and it has developed a remarkable set of propaganda tools that it has aimed at democratic nations. We know that in addition to fake social media accounts, they have attempted to hack voting systems and in other countries even managed to get fake election results posted on government sites. This has already altered US policy, Paul Manafort is under indictment and during his reign the Republican Party's platform quietly dropped opposition to Russia's invasion of the Ukraine.
    We do have some causes for optimism.
    1. First mover advantage. I remember in the 90's this new thing called a 'blog' raised hackles, this site called 'The Drudge Report' caused a lot of trouble with poorly sourced stories and rumors. It's still around today but no one cares. Blogs are a dime a dozen. The Drudge Report took advantage of being among the first on the national scene...they got the credibility of established media outlets without paying the cost of long experience, an army of fact checkers, bending over backwards to get the story right etc. Today the public is more hip.
    2 Numbness. Consider the first ads for Ike or JFK, or the first movies where people jumped out of their seats when they saw a train coming at them. Gimmicks work initially until people get familiar with them, then they become a lot less potent.
    3. Success attracts imitators. A successful marketing strategy is the least stable land to build on. Others follow suit. Armies of tweetbots selling movies, soap, toilet paper will leave very little influence left for Russians.
    Finally, nothing teaches like getting burned and if this administration is anything it is a harsh lesson to Americans in the cost of making the wrong decision. Trump is what you get if you elect a Putin but forget to give him control of the media and culture. IMO nothing illustrates this more than the sad line of 'communications directors' the White House has gone through.
    But the burn isn't just with Trump. We are seeing with Brexit the supporters have vanished, the Russian tweetbot army didn't stick around to explain why the National Health Service didn't suddenly get a lot of extra money that was going to the EU. We can add to this the health care fiasco in the US, the tax 'reform' that might fare a bit better...I won't be cruel and add the 'Wall with Mexico' to the list.

    1. 1- First Mover advantage: Most people use the first mover advantage to refer to the long term advantage which comes from being the first in a space. The way Amazon continues to dominate ecommerce by being the first big player. But I also take you're point there is something to doing something before people adapt, though I don't know that the Drudge report is the best example:

      2- Numbness: This appears closely related to the above, but I wonder if things will take a different trajectory. If you look at the speed with which movies change it's pretty slow, social media ads can change a lot faster.

      3- Attracting imitators: This is where I wonder about 2020 (or 2018) if it was as effective as I think, you're going to have a substantially greater percentage of the 2 billion in total spending than the 81 million you had this time around. I don't hesitate to predict that absent some major event (Facebook being out of business for example) that it FB spending will be more than triple what it was this year come 2020. And that's probably conservative.

    2. has an intereting article on Russian bots and Brexit. An interesting takeaway is just 139 tweets from 29 accounts were retweeted by 268,643 people. So 29 fake accounts moved over a quarter million people to retweet a 'thought' they found worthwhile. I think this leverage is much more important than traditional advertising spend which can be $81M for billions of impressions. The fake accounts were cultivated for a while in order to attract followers and seem like real people. This reminds me of a podcast I heard this morning that talked about the fall of Ashley Madison, the dating site for those looking for affairs. When hackers broke their database and made it all public, it was revealed males far outnumbered females but also many of the few female accounts that existed were really just bots pretending to be conversing with males.. Also I'm not getting any detail around the $81M number for both campaigns combined. Is that all impressions or did Facebook charge them for setting up customized pages?

      I think this clearly indicates what Russia did mattered a lot, but now the cat is somewhat out of the bag. it's not that hard for other people to set up fake twitter accounts to drive their own media buzzes. A similar thing happened with customer reviews when companies realized they could drive their products up by having fake accounts 'review' them. "Search Engine Optimization" where people try to get higher google ratings has been going on for more than a decade now with 'white hat' and 'black hat' operators deploying all types of tricks to figure out and game Google.

      Anyway I suspect next time around we are going to have more citizens skeptical of say a @SouthLoneStar who claims to be a proud Texan who tweets the Kremlin party line about European politics. I also expect him to be joined by other bots pushing other people's interests.

      I think, in other words, what Trump and his allies did mattered but I'm skeptical whether the 'trick' will help much next time around. First Presidential candidate to use TV had an edge, it wasn't long before all of them used TV well, though.

  2. after listening to your podcast I tend to agree that people will adapt. Two things I think we have to remember:

    1. A/B testing is a moving target. Unlike a clinical trial, the results you get from a well designed study on human behavior are temporary. If people, say, shop more when the store plays medium tempo music, there is no guarantee that relationship will hold into the future. Tomorrow stores with fast music may get more sales. Unless algorithms reveal unchanging aspects of human behavior (important point there), the ability to exploit them is temporary and can suddenly backfire. In other words, being a slave to fashion is pretty unstable and nerve racking.

    2. Serendipity is a pretty important value. Back when Barnes & Noble & the late Borders brought in the concept of the big box bookstore, hanging out there was a lot of fun but since you could walk around you could easily run into books outside your preferred areas of reading. Libraries do the same thing. I'm not sure Amazon has figured that out yet but people do like things mixed up. I don't believe the algorithms are just a bottomless pit of more and more partisanship. People do want some stimulation with the different rather than just endless confirmation with what they are familiar with. Do people not try different types of food that seem exotic even in areas of the country where the menus are pretty stable 99% of the time?

    1. 1- I totally agree that A/B testing and marketing are movable targets, this is part of what causes the replication crisis in social science in my opinion. But also this means that when people adapt to one type of ad it's that much faster to try out a different type of ad. Consider how much more viral internet content is than anything that was ever on TV. Part of that is obviously ease of sharing, but part of it has to be that there are so many more potential viral things to begin with.

      2- Maybe, but I haven't seen much evidence of that. All I see is more attempts to double down on what's already working. If you see any evidence of diversification please pass it along because I'd be very interested in it.

    2. #1. Nicholas Christakis had an interesting take...imagine 4 people are tied to each other and blindfolded in an unknown territory. Their task is to find the highest point and stay there. One method might be to take a step in a random direction, if they move up take another, if they move down reverse, is they stay level pick another random direction. Another method would be to allow many steps in the wrong direction. The first method will quickly get the group to the top of a hill, but not necessarily the highest point. The second will map the entire terrain and find the highest point.

      I think A/B testing is like the first method. Fox News might keep so many viewers by swapping one very young, large chested female 'anchor' for another one that's slightly different or by alternating 3 Clinton conspiracy stories to 1 Trump praising story but does that achieve the highest point possible esp. when the highest point can radically change? For that you have to try things that are radically at odds with your AB testing method. In media this is pretty critical because the cost will be your competition may discover that higher point and take it over.

      #2 I think we have a bit of contradiction here. I remember when tacos were 'exotic foreign food' but today it is fast food. On one level there seems less diversity as you'll find the same fast food junk at every exist off every interstate highway in the country...and even the world. I think people have a craving for 'something different' and the market is trying to provide it but there's an equally strong desire to create comfort in sameness which is also being provided. So McDonald's is everywhere, but there's more on its menu than ever was.

  3. I would disagree when you say "On the other hand, the TV and Radio revolution was almost entirely peaceful and we can always hope that even if social media does represent a revolution it will be more similar to the revolution brought on by TV and Radio than the one brought on by the printing press." Particularly in regards to radio. I think it can be argued that the propagation of destructive ideologies, mobilization and coordination of 100,000's warriors and the vast scale of the wars during the last century could not be accomplished without the ability of broadcast radio. Leaders were humanized, orders could be transmitted to the uneducated, no unnatural reading and writing were needed. There is more information in the human voice, less ambiguity, easier to stir the passions with speech and I think the people reacted in bigger and more unified ways. Language barriers became one of the few brakes on this consolidation. Of course, once one person can direct these forces we get WWII scale conflicts and ambitions.

    Television is less clear to me. An audio recording of a massacre would be horrifying, as would a video recording but it seems like there was never a time when people sat around and listened to the horrors of war and yet in the early 60's and 70's it was common to watch shows with footage from ongoing wars. Did this make us lose our taste for large scale war?

    As to our current mix of media, it seems hard to plan around as new people are born, new social platforms get made and fads come and go. I guess I predict fewer large scale efforts as attentions fracture and more outside influence as more people master the, now cheap and ubiquitous, tools. Even though less people will be working on individual projects I expect their results to be world changing. I think anti-fragile is a good way of thinking about how to operate in such a future.

    1. You make a good point, I was probably overstating the case to say that TV and Radio were almost entirely peaceful. Though when you compare the bloodshed which attended the printing press, it was pretty mild, but you make a good point about WWII it's hard to imagine Hitler being quite as successful if not for Radio or Triumph of the Will, nor is it certain that Britain would have made such a stout resistance without Churchill's radio addresses.