Chris Wilson is the current CEO of WPA Intelligence, a conservative research and data analytics company which provides information to an array of clients, mostly political campaigns. Wilson served as the Director of Research Analytics and Digital Strategy for the Cruz for President and is a regular political analyst on both Fox News and MSNBC.
Harvard Political Review: What is WPA Intelligence and what is your mission?
Chris Wilson: We are a Republican polling data analytics company. We do work all over the country and the world. We were involved in the 2019 Canadian elections, for instance. I would consider us to be an innovator in analytics, and certainly the largest on the Republican side in terms of producing polling and analytics for Republican candidates.
HPR: Your technology played a large role in mobilizing voters for Senator Ted Cruz’s (R-Texas) 2016 election campaign. What are some things that you learned from that campaign?
CW: That analytics isn’t everything. We still lost. You can always have the best data, the best operation, and still come up short in a campaign. And I think we had the best candidate, too. Because we did have a large enough budget, we ended up spending $92 million by the time it was all said and done — enough to really innovate, and we tried things that no one had ever done before in politics, whether for a Republican or a Democrat. And that experience now informs everything that we do for all of our candidates, whether they’re running for the Senate, governor’s office, state house, or dogcatcher.
HPR: Given that Democrats are using similar techniques and technologies, how much do you feel that campaign technology will affect the upcoming election?
CW: Time will tell. I think Republicans have wisely combined efforts, through Data Trust and the Republican National Committee data operation, to cooperate with each other to make sure that no matter what happens in a primary, the nominee comes out stronger and with a solid data platform to work from and campaign from. And I think that is an area that we are very far ahead of Democrats. Democrats don’t have anything like Data Trust or the RNC data operation.
And until they do, I think [Democrats] are going to find themselves behind, which is shocking when you consider Daniel Kreiss’s book, Prototype Politics, in which he studies the 2012 election cycle. And in it, his study is looking backwards at the 2004 Bush campaign, how they innovated between 2000-2004 and created micro-targeting, which was a precursor to data analytics. But after 2004, Republicans stopped innovating, and it allowed the Obama campaign to emerge in 2008 running a much more advanced operation. Then, the Obama campaign totally innovated between 2008 and 2012, and in the 2012 election, there was no way Republicans could catch up.
And so now you’ve had … the effort behind the RNC on one side, and the Trump campaign on the other, invest[ing] millions and millions of dollars into data. In fact, if you watch Brad Parscale’s Twitter feed after every rally — that’s Trump’s campaign manager — he tweets out about all the data they’ve collected. He tweets how many people at the rallies were Democrats. At a recent rally in Dallas, I think the data said that more than a third of attendees were Democrats and 20 percent were Hispanics. I mean, it’s just a remarkable amount of data they’re collecting and ingesting into their overall system. It’s going to be very difficult, if not impossible, for anyone to ever match that.
HPR: How is your company working to shape the 2020 campaign environment?
CW: We work with all the party committees. And so from that standpoint, we’re a part of the RNC data operation. And we’ll be working for our individual clients as well. And [it] remains to be seen how many [clients] there are. During our last election cycle we did over 500 races around the country.
HPR: So are you going to be focusing away from the presidential campaign and more on smaller races at the state level?
CW: I think our goal would be to work with the National Republican Congressional Committee to make sure that Republicans hold the Senate and assist individual campaigns to help Republicans take back the House. I think analytics clearly matters at the presidential level. But the smaller races are ones in which you could have a substantial impact.
HPR: Data analytics seems to be focused a lot on persuasion. How does this fit into the “get-out-the-vote” efforts once you get closer to the end of a primary or the end of the general election?
CW: You know, one of the beauties of analytics is that you score every voter on their likelihood to vote and their likelihood to vote for a specific candidate. I think one of the reasons you’ve seen turnout increase to record levels in the last two election cycles is because you see efforts that are data-driven to reach both new voters and inactive voters. In Texas, there are 28 million residents, and there are 15.6 million registered voters. Your typical turnout in an off-year election is about 5.5 million people. And last cycle, in 2016, it was 8.6 million people. That is an increase of almost 2.2 million. Now a lot of that had to do with Beto O’Rourke, but a lot of it had to do with Ted Cruz.
We began an effort early on for both Sen. Cruz and Governor Greg Abbott (R-Texas), in which we identified 2,068,746 people to be exact, turnout targets. By turnout targets, we mean people who we knew would vote for both Abbott and Cruz, but did not currently plan to vote.
We began targeting those voters the day after the primary, and we communicated with them digitally seven times a week from [then on] until the election. And of those 2,068,746 people, we identified early on, all but about 200,000 of them ended up voting.
Beto O’Rourke used to say, “Texas isn’t a red state. It’s a nonvoting state.” I think Texas has proved that it is a red state, even when it votes in full numbers. And I think you’ll see those numbers exceeded in 2020. I think you’ll see Texas set a record for election turnout. Then you see turnout reach 10.4 million people. And that’s of the 15.6 million people that are currently registered to vote. And you could see another million people come on to the voting rolls.
HPR: What does the timeline for data analytics look like? Is it important that you start early?
CW: I don’t know that there’s a perfect timeline. I think when it comes to politics, the only thing that you can’t get more of is time. I always like to be able to start as early as possible because I want to communicate as soon as possible. You can always get more money [and] you can always get more volunteers, but you can never get more time. And so every day that you lose, [every day] that you’re not communicating with voters, I think is a day you’re not going to get back. So I’d rather have plenty of information, and [have it] as early as I can.
HPR: How do you address the privacy issues associated with data analytics?
CW: The value of analytics is your ability to communicate directly with people. I think the reason why you’ve seen turnout increase so much over the last few election cycles is that voting analytics allows people to talk about issues that they specifically care about and not issues that only 10 percent of swing voters care about. When it comes to privacy, I think that efforts have been undertaken by Twitter and Facebook to make sure that voters understand how their information is being used and who is communicating with them.
That’s an important step, and there’s probably more that can be done. I think both Facebook and Twitter and other platforms acknowledge that. But really, the value of personalization to the individual, whether it’s a song you listen to on Spotify or Amazon telling you that you’re almost out of toilet paper, is the ability to make predictions about for you. It makes our lives easier and better. And I think politics is at least as important as what song we listened to or what toilet paper we buy.
HPR: What does the future of this technology look like? And how are you continuing to improve and adapt?
CW: I would have never predicted we’d be able to do what we can do today. In 2010, we built that first predictive model for Sen. Mike Lee (R-Utah). It took us a week to run the model, and that was only 300,000 voters with 400 pieces of data. In Texas today, we can score the entire 15.6 million voter file with over 200,000 pieces of first, second and third party data for almost every voter in the state. And we can do that in a matter of hours. And that’s because of two things: Moore’s Law, which states that computing power keeps becoming more and more powerful, and the inexpensive amount of storage. I had to buy another server to be able to run that first model for Mike Lee, whereas now we can just do everything through Amazon Web Service or Microsoft, or multiple other cloud-based services.
I think as artificial intelligence improves and machine learning algorithms become more sophisticated, we’ll be able to predict anything. And I believe that we’ll be there soon, and we’ll be able to communicate with people even more directly than we [can] now.
Image Credit: The Oklahoman / Paul Hellstern