Sorting through the Facebook-Russia-Trump Advertising Story
We learned from the Washington Post last week that a Kremlin-backed Russian firm spent at least $100,000 on Facebook ads to influence the 2016 election in favor of Donald Trump. Writing in The Intercept, Sam Biddle argues that it is time to make Mark Zuckerberg testify. there’s still a lot we don’t know at this point. Biddle raises the following questions:
What was the content of the Russian-backed ads in question?
How many people saw these ads? How many people clicked them?
What were the Facebook pages associated with the ads? How many members did they have?
What specific targeting criteria (race, age, and most importantly, location) did the Russian ads choose?
Some of these questions matter more than others. This Russia-Facebook story is legitimately explosive, but not for the reasons you might immediately suspect.
I believe three things can be simultaneously true:
- The Facebook-Russia-Trump story is a big deal. It may be further evidence of direct coordination between members of the Trump campaign and foreign propaganda outlets. Mueller should investigate this fully, including what (if any) role intermediaries like Cambridge Analytica played in helping Russian troll farms to segment and target their messages.
- This story further underlines some troubling realities about Facebook today. Facebook is enormously influential in driving web traffic. Engineering decisions at Facebook have massive social consequences, shaping both what news reaches citizens and shaping the incentives for which news outlets will thrive in 2017 and beyond. Facebook effectively has monopoly power in this space, and goes pretty much unregulated. Unregulated communications monopoly power ain’t great, even when the monopolists are well-meaning.
- The $100,000 ad buy itself had virtually no impact on the outcome of the 2016 election. Whether those ads reached 30 million or 70 million potential voters doesn’t particularly matter. How many people clicked on the ads doesn’t particularly matter. Targeting criteria and message content doesn’t particularly matter. Digital advertisements don’t swing national Presidential elections.
These three points can be simultaneously true. There’s no inherent conflict between them, no need to chose a side among them. But they are often placed in stark contrast to one another. Either you believe microtargeted computational propaganda duped the masses or you believe all this “Russia stuff” is overblown. Either you think Facebook is a new information monopoly or you think social media is overrated.
I’ve particularly noticed a habit among researchers who are genuinely (and rightly) concerned about #2 to overinflate the threat and potential impact of Facebook advertising. The same thing happened during the emotional-contagion controversy and during the trending topics controversy.* If your focus is on reigning in Facebook’s unbridled power, it is easy to adopt a maximalist interpretation of any breach or misstep.
But the result can look kind of silly, and it puts us in a false intellectual bind. $100,000 is really small potatoes in a presidential race. U.S. presidential elections are a multi-billion dollar industry. The Trump campaign itself spent tens of millions of dollars on Facebook advertisements. These Russian Facebook ads are not the videotape from The Ring(…if you watch it, seven days later you vote Trump!).
The literature on political advertising and persuasion is pretty clear on this front: Big tv ad buys produce short-term, marginal shifts in voter intention. Those changes vanish after a few days. Most people don’t notice political ads, the ads that we see are unlikely to change our opinion of the candidates, and when our opinions do change, they often revert back over time. Even with advances in microtargeting, there is no evidence to suggest that Facebook ads are more persuasive than television ads.
This is a hard-won lesson from the field of political communication. Political propaganda does not function like a hypodermic needle, injecting new opinions into the populace. Successful political persuasion, simply put, is really hard. When we conflate exposure-to-propaganda with behavior change, we are taking the propagandists marketing brochures far too seriously.
Many of my peers in the political communication research community have responded to the wave of Facebook stories (both about fake news and psychographic targeting) with dismissive eye-rolls. We know this stuff doesn’t work very well. How many times do we really have to re-teach this literature?!? But in the midst of those eye-rolls, we run the risk of overlooking the substantive importance of these stories. What matters isn’t the direct impact of the Russian advertising. What matters is the chain-of-data-custody that led to the ad buys.
To draw an analogy, there’s no reason to think that the 1972 Watergate break-in at the DNC headquarters changed the outcome of the 1972 election. The break-in mattered because it was part of an illegal attempt to influence the election. And it kept mattering because of who was involved in orchestrating and covering up the break-in. If it had turned out that “the plumbers” were actually just a bunch of drunk White House interns, then the scandal itself would’ve fizzled.
The immediate reason that this $100,000 Facebook ad buy matters is that it didn’t come from nowhere. Maybe Russia threw money at these advertisements just because Putin likes to mess with democracies, and he has oil money to burn. (Except no, per Jim Rutenberg, this is Russia’s new strategy for disrupting democracies through information warfare.) But if it turns out, for instance, that Brad Parscale or some of his peers at Cambridge Analytica were sharing targeting and voter modeling data with Russian cutouts, then we have a clear-cut criminal conspiracy here.
It doesn’t matter if the Russian Facebook ads included microtargeted pixie dust or not. It doesn’t matter whether they were persuasive or changed the outcome of the election. You don’t have to believe that computational propaganda is radically effective in order to care about this scandal, just as you didn’t have to believe that the Watergate break-in fundamentally swung the 1972 election in order to care about Nixon’s abuses of power.
It also doesn’t matter whether you think Facebook is a communictions monopoly in dire need of regulation or a take-it-or-leave-it social media platform.
There was a strategic, intentional campaign by foreign actors to influence the 2016 Presidential election in favor of Donald Trump. There is a trail of evidence suggesting that senior members of the Trump campaign facilitated this illegal influence campaign. That is a crime, and it has been accompanied by a massive cover-up.
So yes, Mark Zuckerberg should testify. His testimony should provide a window into the essential, unregulated role that Facebook plays as an information intermediary. But more importantly, Zuckerberg’s company should collaborate fully with Mueller’s investigation. Because the central question here isn’t just “what specific targeting criteria did the Russian ads use.” The central question is where did this targeting criteria come from? Zuckerberg himself cannot answer that question. But he and his company can help move the investigation forward.
*And hey, I teach strategic political communication for a living. I get it. Public attention is rarely pointed at controversies over information infrastructure. Use it or lose it in rare moments like this one.