Over recent weeks we’ve learned much more about how Russian operatives used Facebook to support Donald Trump, attack Hillary Clinton and spread conspiracy theories pumped up the heat of the 2016 campaign. One big question has been: how effectively did they target those messages, given Facebook’s vast ability to target messages? And if they did target their messages to areas of particular Democratic weakness in states like Pennsylvania, Wisconsin and Michigan, how were they able to do that? Where did they get the data to drive the effort?
One possibility is obvious: Maybe the Trump campaign gave the Russians access to their data and voter files. To date, there’s at least no public evidence that this happened.
But maybe it didn’t have to.
What I’m about to relay is not new information. But I’m not sure it’s been considered in the context of the new information about Facebook and other digital campaigns apparently run by the Russian disruption campaign which we’ve only learned about quite recently.
Back in May The Wall Street Journal ran a fascinating article about a Florida political operative named Aaron Nevins. When news reports surfaced identifying the online persona Guccifer 2.0 as the one who had the files hacked from the DNC, Aarons essentially cold-emailed Guccifer 2.0 and asked if he had any material to share on Florida. That was on August 12th, 2016. It turns out he did. A lot.
10 days later Guccifer 2.0 sent Nevins 2.5 gigabytes of data from the Democratic Congressional Campaign Committee. Note that that is different from the Democratic National Committee (DNC), the hack we’re most familiar with. The Journal described the documents like this …
DCCC documents sent to Mr. Nevins analyzed specific Florida districts, showing how many people were dependable Democratic voters, how many were likely Democratic voters but needed a nudge, how many were frequent voters but not committed, and how many were core Republican voters—the kind of data strategists use in planning ad buys and other tactics.
The Journal reviewed these documents as well as Democratic voter analyses also sent to Mr. Nevins about congressional districts in Kentucky, Pennsylvania, Texas, Virginia and West Virginia.
Nevins posted some of that data on his anonymous Florida politics blog. Guccifer 2.0 in turn sent that link to Roger Stone. There’s a whole other mystery about what happened to that data from there, who may have used it, what Stone did with it and so forth. But let’s set that aside for the moment. Nevins just asked for whatever “Florida based information” Guccifer 2.0 might have. He had a lot. There’s every reason to believe that Guccifer 2.0 had comparable data for other states and that he had data from multiple Democratic campaign committees. At least he had information from the DCCC and the DNC.
US intelligence believes Guccifer 2.0 is a fictive persona created by Russian military intelligence. The Facebook campaigns appear to have been run out of a St. Petersburg, Russia troll farm called the Internet Research Agency (IRA). The IRA is nominally a privately owned operation. But it seems clearly to work on behalf of the Russian government, even if it is technically independent from it. In any case, the big picture should be clear. If the Russian election disruption campaign needed election and voter data to effectively target its digital campaigns, they seem to have had a lot of it, precisely the kind of detailed data on strong partisans and more marginal voters that would be key to directing such an effort.
Certainly, whoever was calling the shots on the Russian side might still have needed assistance making sense of this material. At a minimum they’d need a reader of English and some basics of US geography. But far from needing guidance from the Trump campaign, the Russian hackers likely had – at least in some respects – better data than the Trumpers did. At a minimum, they had Democratic-side data which could have been used to great effect on its own or created an even fuller picture if married to the data Republicans and the Trump campaign had.
