Let's start by talking about the movable middle. And to do this, let's talk again about politics. So this is information you probably know already, but I think it's important to point out. Every election cycle political campaigns spend huge amounts of money. In 2016, for example more than $6 billion was spent on presidential and congressional elections. Some is spent on staff, some is spent on food, and some is spent on transportation. But most of it's essentially spent on persuasion. Things like direct mail and phone banking and door-to-door canvassing. Things that are trying to convince voters to go one way or the other. There is some evidence that this money is well spent. When political scientists looked across dozens of studies on primaries and ballot measures, they found a clear pattern. Advertising and campaign contact worked. Things like direct mail or door-to-door canvassing influence both who voters voted for as well as how they evaluated different candidates. But when these same political scientists analyze general elections, they found something else. One might expect the same results, right just like primaries, races to decide who would be president or hold a senate seat involve contacting voters, trying to persuade, advertising and so on. But across dozens of studies of these more general elections, the scientists found that the same persuasive tools had zero effect. Just to be sure they conducted even a few more tests, they ran new experiments, studying thousands of people. And while this approach increased statistical precision, it netted the same thing. No effect, why? Why does the same things, persuasive campaigns, work in things like primaries and ballot measures, but not work in general election? And while the answer lies in the difference between these has two types of voting, both involve multiple candidates, competing against one versus another, lying out there stands in various issues and trying to win over votes. But rather than run against members of their own party, as in primaries, general elections, usually involve running against someone of the opposing party. Someone on the opposite side. And so while primaries involved deciding between two candidates who staked out positions we could call on the same side of that field, the same football field, general elections involve someone who's on one side of the field and someone who is on a completely different half. Not only are the two opinions far apart, but one is likely in the zone of acceptance, while the other is likely in the region of rejection. And so consequently changing minds in these situations become a lot harder. It's one thing to get a Democrat to move a little bit, to support a Democratic candidate that slightly more liberal or slightly more conservative than them. But getting them to support a Republican is going to be a lot more challenging. And this is particularly true given the strength of party affiliations. Feeling strong about an issue or domain, changes that zone of acceptance. The more we care about something the smaller the zone of things we're willing to consider. Smaller zone of acceptance and wider region of reject. And this is part of why changing political views is so challenging. You're not just trying to change positions slightly, you're trying to get people to switch sides. And not on just any issue, one they feel really strongly about, and they're unlikely to take alternate perspectives. It's like asking a Red Sox fan to start rooting for the Yankees or a Coke drinker to switch to Pepsi. It's not the easiest task. And so what do we do in a situation like this? Do we just give up? Well, not quite. Because the election study found a silver lining. One place where candidates changed minds in general elections, even when change seemed quite difficult, and that was by finding something called the movable middle. Because in politics, smart campaigns don't try to change every mind. They focus on those swing voters who are open to facts and arguments. Undecideds or pockets of people who, given the candidate, circumstances or issues, are more receptive to being swayed in the first place. People who have a larger zone of acceptance or a zone that overlaps more with the candidate's positions. And rather than using the same arguments on everyone, smart change agents use a more surgical approach. They target people with specific messages that are most useful for them. In 2008, for example, a Democrat named Jeff Merkley was running against an incumbent named Gordon Smith, for the US Senate in Oregon. Smith is quite popular, generally viewed as a moderate, and so the race was quite competitive. Researchers were interested in seeing where they could change minds. Where they could get people who would have voted for that incumbent Republican to vote for the Democratic challenger. But instead of carpet bombing every one with the same appeal, they worked to find people who might be already predisposed to switching. Voters who, for whatever reason, might be more willing to change their minds. They started with a wedge, an issue where the incumbent was out of step with at least some of his constituents. After sifting through various possibilities, they landed on abortion. Oregonians tended to be pro-choice,and the incumbent Republican was not. Even better, the challenger was on only a handful of Senate candidates that the National Association for the Repeal of Abortion Laws, or NARAL, was endorsing that year. Then once they found that wedge, the researchers identified pro-choice voters and tried to persuade them. Early in the election cycle, they had sent out various mailings and so they had some sense, through phone calls and other ways, of who might care about the issue. So they targeted them. Would that approach work on all voters? Definitely not. The economy dominated the election and abortion was not the most important thing for everyone. But for some set of people, this particular message was quite effective. And by going after the movable middle, the campaign shifted voting by almost 10%. And that Challenger won. And so when dealing with issues where people feel strongly, we need to start with the movable middle. Individuals who, by virtue of their existing position, are more likely to shift, because they're not so far away to begin with. One way is to look for what I'll call behavioral residue. Clues that indicate contradicting opinions or a willingness to change. In the political context for example, Democrats who support gun rights, or Republicans who signed petitions supporting environmental reform. In the business context, for example, going on social media and looking for consumers who complained against a competitor. Those people are more willing to switch in the first place, they're already in that movable middle, and they're easier to target. And even when data isn't available, try testing and learning. Take a sample of people, test a particular approach, and report key characteristics on various dimensions. Using that can help us identify subgroups that can determine what types of people might be good to go after more broadly. Trying to get a new product to take off rather than try to convince everyone, find a subgroup. Find that set of people that already needs it and go after them. Venture capitalist often talk about products and services as vitamins or painkillers. Nice-to-haves, vitamins that can be put off until later, or need-to-haves, those painkillers that people can't live without, right? Rather than going after everyone, we need to start by finding the people that see an offering as a painkiller. If we're new startup, for example, who are those individuals we don't need to persuade, but if they just knew we existed, would be already interested in what we have to do. Locating these potential users who need the offering and can't wait to sign up. Find these folks, not only would they change their own minds, but by doing so they'll change minds around them, as well.