Below you will find the highlights of a news story well worth reading that discusses Facebook’s recnet Cambridge Analytica Scandal. How will this impact marketing online? It remains to be seen.
For a few days there, you could hardly look at political news without seeing the words Cambridge Analytica. As soon as word broke that Facebook had banned the UK-based data analytics firm from its platform on March 16, stories popped up connecting it to everything from Russian officials to campaign dirty tricks around the world.
Facebook cast Cambridge into the wilderness because the company violated Facebook’s terms of service when it “scraped” data from some 50 million Facebook users via a Facebook app in 2014. Many other advertisers — including the Obama campaign — also obtained user data via apps in the same period (including information on app users’ Facebook friends), but CA apparently ran afoul of the rules in the ways it shared the data.
The scraped Facebook profiles served as the foundation for CA’s “pychographic” targeting, which they pitched to potential clients as having the power to tailor advertising and outreach based on people’s deep personal characteristics. Best known for working with the Trump campaign in 2016, the company also consulted with the campaigns of Ted Cruz and Ben Carson — along with many others around the world.
What does this wave of revelations mean for political professionals? Let’s break it down:
CA lived and died by a sword made of data, and its abuses have inspired lawmakers and advocates in the U.S. and U.K. to call for stricter data regulations or better enforcement of existing ones. The practical results will depend on exactly what laws or rules they manage to pass, particularly after the story drops from the headlines and people aren’t paying as much attention.
A Facebook movement sprang up almost as soon as the news of CA’s banning broke. Less than two weeks later, it had already enlisted celebrities like Cher and tech leaders like Elon Musk, who deleted the Facebook pages of Tesla and SpaceX.
The company itself could impose new restrictions on advertisers, but it can’t cut their options too much without killing the company’s revenue engine. Campaign communicators should watch for changes. Still, I don’t see any reason to reduce your 2018 Facebook budget unless something more dramatic happens.
Now we’re getting to the dangerous part, at least if you’re a CA employee or client. For a start, the staff may have coordinated their work for the Trump campaign with their work with other clients, including pro-Trump PACs. That’s a big no-no under U.S. campaign finance rules, though with the FEC deadlocked, the company may escape consequences — unless Robert Mueller’s investigation takes up the thread.
Domestic coordination is just the start, since it seems that CA used foreign nationals for the bulk of its work with American clients. Once again, that’s a practice banned under U.S. law, if the staff in question made substantive strategic decisions for the campaign. And then there are those whispers about the Russians. If they pan out, all bets are off.
At the very least, the CA scandal (CambridgeGate? AnalyticaGate?) gives extra impetus to the drive to include disclosure language in online political advertising, though new rules would almost certainly not take effect this election cycle. This year campaigns should expect extra scrutiny on any activity that might look like coordination, though.
Was Cambridge Analytica Selling Snake Oil?
Behind the broader CA discussion hovers a basic question: did their psychographic targeting live up to the hype? Brad Parscale has long claimed that his team didn’t use Cambridge’s data when they built Trump’s comprehensive Facebook outreach program, and it appears that he treated CA as one of many vendors competing to demonstrate the utility of their tools.
In any case, psychographic targeting is nothing new — vendors have been pitching some variation of the idea for years. Psychographic profiling can help campaigns establish their initial target models, but data coming in from direct voter contact should supersede the models quickly. Who needs an estimate of who might respond to a given message once you can watch what they click on? Once voters tell you what they care about, you don’t need to guess.
A good rule of thumb? If it sounds like magic, it’s might just be vapor, no matter how fancy the Powerpoint. Pro tip: vaporware rarely wins elections.