The Case Against Microtargeting

What a victory! This is what the start of a durable majority looks like. If we are able to take this momentum and deliver for the American people while running good campaigns I believe we will be able to improve a lot of people’s daily life. One thing we need to keep an eye on though as we move forward is not overestimating our own mandate from the American populous. We won a lot of difficult races and it is essential that we start thinking now about what our vision for the future will be and what strategies we will use to share it the American people so that we can hold our gains in 2028 and beyond.

One major aspect of Democratic strategy since the 2000s Democratic campaigns has been targeting contact specific voters based on a variety of characteristics such as their age, gender or scores that predict who is potentially persuadable. These systems really scaled up in 2012 to help Obama get reelected and led to our current system today which allows for targeting based on a wide array of characteristics or scores that estimate people’s likelihoods to interact with the political system.

In a lot of ways this is good! I heard last week the campaign manager for a senate campaign had their door knocked by a volunteer supporting their own candidate. It is not a good use of our volunteer’s time to talk to folks that are already in our corner. Ideally a good score would mark that campaign manager as a slam dunk for the campaign that doesn’t need to be contacted and the volunteer would know to skip that door. Similarly, there are plenty of people out there who simply do not want to hear from our side of the aisle. Avoiding them can improve a volunteer’s experience and avoid turning out Republican voters. Targeting using modeled characteristics is a great way to do this.

However, this approach can be taken too far and have corrosive impacts on our Democracy. For instance, suppose I was running a candidate for president and trying to target very nitty gritty policy pitches to the small groups effected by each one. This sounds really good in theory but the internet makes it very easy for messaging to escape its target audience, and the campaign focus can feel very specific rather than a universal vision. Instead of having a solution for each problem you risk not having a universal message that can appeal to the broad population.

Moreover, it can have negative impacts on the functioning of an organizing program. These programs are how we contact voters, usually face to face or over the phone generally led by volunteers. If you are targeting only a small group of people deemed best by the model you risk risk contacting the same people multiple times instead of lots of people a few times which is much better given the impact of diminishing returns. Small target groups also degrade a volunteer experience which will involve spending more time driving from place to place instead of talking to potential voters.

One classic model is doing persuasion in the summer and turnout in the fall. Many campaigns will talk to folk who they think will vote but may not be supporting them already until labor day and then focus all of the contact after that on voters that have middling likelihoods of turning out that we are confident would support us if they made it to the polls. There is evidence behind this showing that you can turn out voters when they are able to vote. However, if your path to victory involves persuasion (which it must in many parts of the country we won yesterday) and the impact of persuasion falls off over time, by running your organizing in these two distinct phases you end up with smaller lists of people to talk to and leave out persuadable voters when their decision is most important.

This is why I encourage campaigns to think about their electoral strategies in terms of the broad electorate. While I definitely use scores and love to think about the contact attempt that will have the most impact, I generally push places I work to target broad swaths of people that maximize numbers of contacts to folks who don’t already consistently support us. In order to defend many of the places we won last night we need to have hard conversations, build our electorate and go the extra mile to win voters over. I am looking forward to continue doing that in the 2028 cycle wherever I go next.