In the wake of the 2016 U.S. presidential election, every politician, party, and activist group realized that tech could be an even more powerful tool for suasion. What are the latest tools for online organizing and how will they change the presidential election?
The following transcript has been lightly edited and condensed for ease of reading.
Nicholas Thompson: We have Lyel Resner, he is the founder of Swayable, a fellow at the MIT Civic Data Lab, an adjunct professor of social innovation at the Berkeley Center for Innovation at NYU. Pinky also ran tech on the Bernie Sanders campaign in 2016, which means possibly working on the Bernie Sanders campaign and the ACLU is the only two times there have been liberal organizations that were well-funded in the history of American civic activism.
All right, let’s get going on this panel because I’m super interested in this because I’m a journalist and every single election cycle my whole life, well, since 2000, I’ve had to write about the most innovative campaign. I remember I wrote about John McCain, he put his URL on the bus. I wrote about Howard Dean. I loved the 2016 campaign because the one that ended up having the better digital strategy was the one where the digital strategy was run by the sys admin at the golf club, not the CEO of Google. That was kind of amazing.
So let’s talk about what you’re doing at the ACLU and tell me about the digital technological strategies you have that are awesome, cool, and that will blow the minds of all of these people right here.
Pinky Weitzman: Thank you so much.
Thompson: Shake them up.
Weitzman: Yeah, the ACLU is—it’s been quite a ride. I’ve been there six and a half years. The last two and a half have felt like about 20 years in the Trump era. We’re a nonpartisan organization, I’m going to start by saying that, but obviously we have some unprecedented challenges that we’re facing.
Thompson: You have some differences with the Administration.
Weitzman: We certainly do. We have differences with every administration, to be fair.
Thompson: Yes, which is one of your great virtues.
Weitzman: We sued the Obama Administration many times, particularly over the use of drones, targeted killing. But we have, you know, nearly quadrupled in size of our membership base and we have 4.4 million social followers now and we need to make sure that we are doing everything we can to use the right digital tools and messages to make sure that we’re giving them timely and meaningful opportunities for activism, and of course to throw us some money now and again.
So we do use a whole variety of tools in order to make sure that we’re using our donor dollars in the most efficient and intelligent way possible. We use tools like Swayable, which my co-panelist, Lyel, is going to talk about shortly. We’re about to switch to an analytics tool called Heap. We have a unique situation at the ACLU with our privacy policy in that we will not hand over any PII to any non-trusted third party vendor. That rules out Facebook, right? That rules out us using any Facebook pixel on our website, Google conversion tracking, that type of thing, so we have to use tools that will play nicely with our privacy statement. CrowdTangle, Spike, which has a tool called NewsWhip, ActionKit, Relay for our peer-to-peer texting. We also have Revere suite of tools, which is broadcast texts and has a calling platform that allows us to quickly set up click-to-call actions for making phone calls to elected representatives. That also ties in with a broadcast SMS on the backend, allows us to grow that list when people make phone calls. They’ll get a kicker SMS that says, “Hey, thanks so much for calling your senator, would you like to join us and get real time action updates via text?”
Thompson: So you’re using all these tools to find new people.
Weitzman: Find new people.
Thompson: But you can’t use location targeting because you’re the ACLU, I’m so sorry. And you’re finding new people, you’re making it easier for them to call Congress, state legislators.
Weitzman: Yes, to take all the different levels of actions.
Thompson: And then once they’ve called, you’re figuring out how to get money from them. Or you’re giving them the opportunity to give you money.
Weitzman: Yeah, I mean, there’s definitely an engagement letter there and I wouldn’t necessarily put donation above showing up at an event, for example, or volunteering.
Thompson: Is there something you’ve got that you didn’t have when you were working for Bernie? Something that’s come out in the last two years, you’re like, “That’s good.”
Weitzman: There are things I will say that—
Thompson: Actually, let me kick it to Lyel. Oh, wait, you have a thing called Swayable that she uses. It didn’t exist when she worked for Bernie. What’s Swayable?
Lyel Resner: Well, I should say, Swayable is just one tool of many, I think, that Pinky mentioned that—there’s been I think a veritable explosion of tools and technologies that have popped up over the course of the last election cycle that aim to promote healthy democracies, peaceful societies, advance human rights. Some of them have kind of explicitly political bends, others are maybe more ancillary and everything from organizing or voter engagement, voter contact, to actual campaign organizing, CRMs. There’s a giant billion dollar industry, kind of even pre-2016, that drives kind of political tech.
So Swayable came in, in a space, in political parlance, appropriately, largely intuitively, is called persuasion which is are you able to actually convince people to either come to your side or take a certain action. So we are a platform that measures how content changes people’s minds. So if you are going to do a giant kind of paid media campaign, six figures, seven figures, even eight, we can tell you pretty much overnight with a scientific degree of precision, video A is 3.7 times more persuasive than video B with this target constituency that you care about, white suburban moms, moderate conservatives. So for your million dollar spend, you can either have 3.7 million dollars of impact or not. Or you could have a million dollars of impact for a quarter of the price.
Thompson: So I decide I’m going to spend a million dollars to make people think that Nick Thompson is a good moderator. So I sign up on your platform, I give you my email, and then you take a bunch of clips of this panel and you show them to focus groups and the focus groups thumbs up, thumbs down?
Resner: It’s basically like a hyper-efficient digital randomized control trial. So we borrow methodology from best practice. We’re kind of a team of former scientists/digital media people and so we kind of recruit thousands of people online, divide them into statistically equivalent groups, give some of them a treatment, some of them a control, and kind of create an index and do some math and kind of out pops the results.
Thompson: And then you tell me this video works, this video doesn’t work, this video works if you’re targeting suburban moms, this video works if you’re—
Weitzman: And in fact, the powerful thing that we found was the ability to find out how much backlash was happening. We at the ACLU might watch a video and think, “This is going to work great, this is going to change hearts and minds.” We found out that, when we tested on the Swayable platform, that actually we were further entrenching voters against a particular candidate. In that case, that was Kris Kobach, running for governor in Kansas. So it was actually hurting us more than helping us among the target that we were going after, which was low-propensity voters.
Thompson: Yes, apologies to the political consultants in the room, but it turns out about 25% of all content that we’ve tested actually has a backlash effect. It turns out that the rules—
Weitzman: And you can’t guess it.
Thompson: So you’re spending money to push people in the opposite direction.
Weitzman: That’s right.
Thompson: It turns out liberals and conservatives tend to respond to different types of messaging, even around the same issue.
Weitzman: So if you’re only focus grouping your content among your peers at your single organization, you’re not going to actually find out what the real impact is going to be, what the real level of persuasion or backlash is going to be. You just can’t, it’s too far inside the bubble.
Thompson: So you guys both think Swayable works, right?
Weitzman: Yes.
Resner: Yes.
Thompson: And you’ll only work with people you agree with, right?
Resner: So we only work with folks—we built it in the wake of the election in 2016 to support folks that we thought were doing good work and protecting things we thought were important. So we are a public benefit corporation, which gives us—is a type of organizational structure that gives us a lot of flexibility in terms of who we work with and who we don’t. And there are definitely people that we will not work with. In general, I guess kind of technically we’re non-partisan, in practice, we’re—you might draw a different conclusion. But we’re very happy to work with folks like Pinky at the ACLU and other folks that we think are promoting healthy democracies, peaceful societies, fighting bigotry, racism, homophobia, misogyny, and so on.
Thompson: And you mentioned the word persuasion when you were talking just a minute ago, which reminded me of the Persuasive Technology Lab at Stanford which famously is where the ideas of Instagram and addictive technology in some ways evolved from. Do you think your product is at all creepy?
Resner: Yeah, great question. Yes. It’s terrifying. We were cognizant from the outset, I mean, for folks that know a bit about machine learning, if you have—in the edge case that we were successful and we knew we were building a giant repository of how content changes certain types of people’s minds, which with good machine learning, which we have, is a mind control device. So we were cognizant of that from the outset and tried to be really intentional about both organizational and product design decisions that we could make in order to ensure integrity of purpose.
I think that set of deliberations which is ongoing and some of the things that we struggle the most with every single day is I think part of a larger movement in tech around really thinking more critically about kind of intention and consequence, everything from—and I think we’re seeing that manifest in everything from consumer backlash and kind of growing regulatory pressure to this fascinating way of worker tech activism that we’re having.
Thompson: But let’s leave that for a minute. So let’s say that, I don’t know, I’m a dictator, I’m running some country, I hear about Swayable, I’m like, “This is great.” So I hire some people, I steal your IP, I build some software, I get my messaging, it’s extremely effective. How do you feel about that?
Resner: I think the reality is it’s already being done. Like there are already people who—
Thompson: So you’re mind control on the good side and you know there’s mind control on the bad side and you’re just trying to fight? That’s where we’re going? This sounds awesome.
Resner: I wouldn’t—
Thompson: Pinky, help.
Resner: I mean, I wouldn’t say—there’s a veritable digital arms race, I think, between people who kind of believe in healthy democracies and people who explicitly don’t and the calculus that we made was this was something that would be a really useful tool for people fighting towards the things that we thought were important.
Thompson: And what is the special sauce? Because the idea of showing videos to a smartly selected bunch of focus groups is not a new idea. What is the special sauce that you’ve got there?
Resner: Well, there a couple different things. Math is helpful and it’s a digital approach.
Weitzman: Special sauce, math.
Resner: So we can recruit thousands of people overnight to participate, effectively overnight to participate in these types of tests, yes.
Thompson: And Pinky, when you sign up with somebody like Swayable, you are the ACLU, you can’t invade people’s privacies, how do you audit them? Like what do you—
Weitzman: Oh, how do audit the vendor?
Thompson: Yes.
Weitzman: We have an extremely rigorous process for assessing new vendors. We look at every way that they would touch constituent PII and we consider even something like an IP address to be part of a constituent’s PII. And so yeah, we put folks through the ringer. And, again, the big tech companies simply don’t meet our trusted vendor standard and we have to be extremely careful about the ways in which we use tools that don’t meet our standards.
Thompson: Okay, so Swayable met your standards, you signed them up, you’ve been happy with them at the ACLU.
Weitzman: Yes, I think we were one of their earliest pilot testers.
Resner: Yes, one of our first clients, yes.
Thompson: What do you think is coming next in the era? Like what is the technology you want for reaching voters—not voters, reaching citizens, persuading citizens, changing democracy, that you don’t have now?
Weitzman: I would say really democratizing the way that we are getting in front of, not just voters, but participants in our democracy writ large. Giving folks tools to have conversations about what’s important, why civil rights and civil liberties are important. Some of these new canvasing tools are really exciting to me, like AOC built a new tool that is called Reach, also a PBC, I think?
Thompson: Yes, I believe so.
Weitzman: That is a really interesting way to make sure that we’re not just doing this old fashioned model of reaching people through campaign offices, sending out people to go knock doors, right? What about in a bar and what about on the subway? And so I think that there are really important ways in which we’re going to be reaching new audiences. We know that it’s no longer sufficient just to mobilize the base and we can see all the ways in which our democracy is a bit more fragile than we thought is was pre-2016. So I think finding ways to have those conversations with one another in a way that is very healthy to our national conversation and using tech to do that through these tools, through these, again, very democratized, effectively canvasing tools but tools that allow us to really broaden the reach of these conversations.
Thompson: So either of you can answer this question, but the fragility of democracy makes me wonder of course about how much the tech platforms have contributed to the fragility of democracy.
Weitzman: Yes, how about that.
Thompson: And my argument would be, yeah, sure, of course they have, and in part they’ve done that by amplifying extremes, and we all know what spreads on social media and what doesn’t. But maybe one of you disagrees or both of you disagree. Do you think that technology so far has made our democracy less stable or more stable?
Resner: I don’t know if it could come down on either side, definitively. I think in a number of ways it’s absolutely made it less stable. I think that’s incontrovertible over the course of the last two or three years in particular, with probably kind of a longer tail feeding into that activity.
Weitzman: And democracies around the world.
Resner: Yes, and democracies around the world.
Weitzman: Hungary and Poland.
Resner: Sure, not just domestically. I do think it’s kind of, again, sort of prompted this larger conversation around intentionality and tech. I think there is, again, this sort of like explosion of tools and technologies that aim to kind of promote healthy democracies and maybe also like advance human rights and protect civil liberties. There’s I think a budding and growing kind of desire amongst tech talent to only work for folks that they think are working in the public interest. I think this surge in worker tech activism across Google and Microsoft and Amazon is not going away and folks are kind of devoting more infrastructure so that some of the movement is sustainable. There’s also folks that are doing really great work at creating tools and frameworks for helping technologists actually be more intentional with how they create. So folks like Doteveryone in the UK, Institute for the Future in collaboration with Omidyar has created a tool that almost functions as an agile event for the technologists in the room where you can surface kind of how—it’s like a risk surfacing tool for thinking about how the technology you create might contribute to surveillance state or the dopamine economy or what its implications are on equity or kind of data, privacy, and security, just kind of broadly how it intersects with the human race.
Thompson: So let’s run that audit on Swayable. So when you think about, okay, how data affect the dopamine economy, the surveillance economy, your answer was, well, we just give it to the good guys, or your answer was we’ve built in X, Y, and Z so that it can never have those negative effects.
Resner: No, I think it’s—that audit is ongoing, right? I think, again, in the DNA of the organization and the DNA of the product, we try to make it so that we would always have space for these conversations and be able to be intentional about it. I don’t think that process is over and it’s something I still think about every day and my cofounders think about every day. So I think it requires vigilance, as I think it requires vigilance on the part of all technologists who take on the privilege and responsibility of creating something.
Weitzman: And vigilance on the part of nonprofit organizations, a little biased here, like the ACLU, who, you know, we work with tech companies often in order to convince them to institute stricter data protection rules so that we can use their products and services or to create privacy-first APIs or, you know, just to push them in the right direction. And then failing that, oftentimes, you know, we are not going to convince Google to change their underlying business model, but we are also interested in working in coalition with companies like Swayable to build privacy protected activism tools. So there are ways to approach this that will ultimately lead in the right direction, which is non-creepy, non-scary toolsets.
Thompson: I am super in favor of non-creepy toolsets. What are the non-effective toolsets, what are the things that everybody is super excited about, like, “Oh, this is great,” and then, “This sucks.” What was it like—if you really have all these vendors walk into the Bernie Sanders campaign and being like, “I can—you’re going to win the election if you just use my tool.” Like what’s the one that actually, “No, that doesn’t work.”
Resner: I think the—
Weitzman: That’s a blur for me, 2016. I’m not sure what happened for that span of months. I didn’t sleep much.
Resner: I mean, I think part of the problem was, at least on the left, that technology infrastructure was lost. There was no institutional memory. Every single presidential election cycle, like a bunch of technology infrastructure was spun up in the course of like nine months and then it went away, often in the form of kind of smaller private companies, a lot of institutional dysfunction, ultimately. And so I think it’s—at least now, maybe it’s less a case of things that people were excited about that aren’t panning out as it is, I think thankfully there are a number of people, again, at least on the left, that are thinking kind of broadly about kind of systems or like an ecosystem type of approach to building infrastructure.
Thompson: Okay, so let’s say every—you know, I think there are 72 people running for the presidency on the Democratic side.
Weitzman: As of an hour ago.
Thompson: If all 72 of them were to use Swayable, what would the campaign look like?
Resner: I couldn’t tell you that.
Thompson: But I mean, what does it look like when two perfectly optimized hyper-targeted ads are running against each other?
Resner: I couldn’t—
Thompson: It’s just more efficient advertising.
Resner: I couldn’t tell you that. It’s just people—I mean, we work with folks that we think are telling the truth or trying to convey an important message and it just will help them understand how to communicate more effectively, more authentically with a certain group of people.
Thompson: Great. Pinky, what is your hope for the way technology and democracy interact in the next couple years? Because we’re almost out of time.
Weitzman: So that’s a small question. I have 36 seconds. Well, yeah, it’s how many months to November 2020?
Resner: I think you gave a great answer before.
Weitzman: What did I say?
Resner: I think folks are kind of leaning towards—or I think realizing the importance of people actually getting out and participating in their democracy, not just on voting day.
Weitzman: Yes, it’s not a spectator sport anymore.
Resner: So tools similar to, I think the way Scott was talking about, think about how to use technology to actually activate people, connect people, bring them out of their homes.
Weitzman: Exactly.
Resner: Those are the technologies I’m most excited about. I hope that continues to move forward.
Weitzman: Yes, I think that’s right. I think the online to offline organizing component of the puzzle is so important now, that we don’t approach the 2020 election like it’s status quo, like it’s just 2016 all over again. Obviously, the stakes are completely different and so if we don’t do a better job of convincing people that it is not simply enough to show up, there’s more significant participation that will be asked of you. And you know, finding ways to reach those folks with, again, really meaningful narratives that convince them of that, using tools, that we know are going to have the most impactful and persuasive outcomes. And put it all together and 2020 starts looking kind of rosy.
Thompson: All right, it’s not enough to just show up at this panel, also go out and activate. And if you see my Swayable campaign, say I’m good moderator. Thank you very much, Lyel and Pinky, it was very fun to be onstage with you.
Participants
