DAVID GREENE, HOST:
Facebook is admitting some mistakes. The company's founder, Mark Zuckerberg, spoke publicly yesterday. This was the company's first public response to allegations that 50 million users may have had their personal data compromised and used in an election. Zuckerberg told CNN that there will be an intensive process to investigate Cambridge Analytica, the firm that gathered the data. But what did the first mistake actually lead to? Well, in 2014, Cambridge Analytica approached a Republican Party strategist named Luke Thompson, asking if he wanted to work with them. He turned them down. He's now vice president of the Republican analytics firm Applecart, and he joins us this morning.
Welcome to the program.
LUKE THOMPSON: Hi. Thanks for having me.
GREENE: So take me back. You were an analytics expert with the National Republican Senatorial Committee when Cambridge Analytica came to you. What were they offering, exactly?
THOMPSON: Well, they were offering a variation on conventional political microtargeting. At that time, they were called SCL Group, and they were making some pretty heavy-handed claims about what they could do that normal analytics couldn't do. Specifically, they asserted that they had what they called psychographic mapping, the ability to differentiate different types of people based on different mental patterns and habits, and that using that information they could gain really, really significant efficiencies and advantages in terms of targeting and messaging over what could be done using more traditional means based on vote behavior, surveys and the consumer data file that's available to all political campaigns.
GREENE: Sounds potentially helpful. Why did you say no?
THOMPSON: Well, there were a couple of things that jumped out at me. First of all, the principles underpinning psychographic mapping didn't really hold up to a lot of scrutiny. The argument that you could, you know, gain these massive efficiencies based on - in politics, based on sort of mindsets, didn't make a whole lot of sense to me when, you know, probably the easiest way to learn what people think about politics is just to ask them about politics as opposed to making the logical leap or the assumption that rigid ways of thinking lead to the way people, you know, will act down the line on political issues they haven't even thought about yet. And then beyond that, they refused to answer basic questions about where they were getting their data. This, you know, in retrospect, was pretty obviously because they didn't want to reveal that they were using a Facebook plug-in to effectively steal first-party data that users were giving over to them under...
GREENE: You were suspicious. I mean, you were both suspicious about whether this is all that helpful, but also suspicious about where this data was coming from.
THOMPSON: Exactly.
GREENE: Can I ask you - I mean, both the campaigns of Ted Cruz and of President Trump did work with the firm. Do you know what they actually did for those campaigns?
THOMPSON: Well, I was not on either the Cruz or the Trump campaigns, but I can tell you what I've observed from outside and what I've heard from people I've talked to. On the Cruz operation, Cambridge Analytica, which was brought in to support both the superPACs backing Ted Cruz, as well as, I believe, the Hard Dollar operation, did a lot of their digital targeting and small-dollar fundraising initially. But they were pushed out at the end of 2015, or at least they stopped being paid by Cruz at the end of 2015. And the digital firm Targeted Victory, which was a byproduct of the Romney campaign of 2012, took over almost a sort of, like, seamless, linear transition in terms of what Ted Cruz's campaign was spending on digital targeting. For the Trump general election effort, which is where Cambridge wound up after the primaries concluded, they were pitched against the RNC's internal data shop. That data shop was backstopped by a firm called Target Point that had two people working in their polling and voter-scoring operation.
GREENE: You said pitched against, like the Republican Party was doing its own work, but then the Trump campaign was also having Cambridge Analytica at least try some of their own work?
THOMPSON: Right. Well, you'll remember, there was a lot of distrust in the wake of the primaries between the sort of Republican establishment and the Trump campaign. And so...
GREENE: Oh, so the Trump campaign was looking to see what they could get from an outside group like this, and that's where...
THOMPSON: No. I think it was more along the lines of, you know, is this group better than what, you know, the sort of wise men at the RNC are capable of doing? Given that Trump had sort of taken the nomination by storm, I think the people in his orbit were not particularly inclined to view more established Republican hands as highly competent. And so they...
GREENE: Well, as far as you know, did Cambridge Analytica, using this Facebook data, influence the election in some way and help Donald Trump get elected?
THOMPSON: As far as I know, no. By the time we got into sort of the meat of the general election, they'd been completely sidelined.
GREENE: But did they help him earlier or, I mean, before the meat of the general election?
THOMPSON: I mean, I'm sure they did some - it looks, based on the sort of expenditures available from the FEC, that they did about $5 million worth of digital ad placement. In the grand scheme of a multibillion-dollar presidential election between Hillary Clinton and Donald Trump, that's a drop in a well.
GREENE: That's nothing. So if there was an influence, it was minimal. Let me just broaden this out. I mean, you are in this line of work.
THOMPSON: Sure.
GREENE: Do you defend it? Is it ethical in general to use personality data to target political advertising?
THOMPSON: Well, I don't use personality data because I think it's nonsense. I don't think it's scientific, and I don't think it's derived from good foundations. But in general, here's what I'll say. I think there's a really clear line that needs to be delineated between first-party data that's derived from users who sign an agreement - sure, nobody reads the terms of service, but you are protected by that terms of service to varying degrees in terms of your privacy, in terms of your data - and using publicly available data or commercially available data to create efficiencies in messaging. I don't think there's anything pernicious about trying to optimize advertising spending. And frankly, I think most Americans would probably prefer to not be blanket targeted with political advertisements, but instead, you know, have them at least to some extent dovetailed to their particular interests or, you know, position in the electorate or likelihood to swing one way or the other. But in terms of taking data that someone has volunteered to give to you and then breaking the terms under which it was given to you for the purposes of targeting, I think that's deeply unethical, and it may very well be illegal.
GREENE: Luke Thompson is the vice president of Applecart, a GOP analytics firm. Thanks for joining us this morning. We really appreciate it.
THOMPSON: Thanks for having me.
(SOUNDBITE OF GARAGE A TROIS' "THE DWARF") Transcript provided by NPR, Copyright NPR.
NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.