Based on the partisan rancor over whether the federal government should create an Election Day holiday, it’s doubtful we will see universal or automatic voter registration soon. Still, it’s heartening to see that voter registration and voter participation increased sharply in the last midterm election cycle–to a fifty-year high, according to NPR.

Here at Accurate Append, we’re always curious about how technology can be used to reach both voters and nonvoters.

The data on unregistered voters send a lot of mixed signals. On the one hand, “43 percent of the unregistered said nothing would motivate them to register” according to the recent Pew Trusts survey and analysis on unregistered voters. This is disturbing news by any metric.  On the other hand, unregistered and “occasional” voters are actually not universally apathetic. The Pew Survey also reported that small but consistent percentages of unregistered and “semifrequent” voters self-report having worked to solve issues in their community, done unpaid volunteer work, and attended community meetings. Some (only six percent) had even donated money to candidates they didn’t end up voting for! Assuming those folks aren’t dogmatic about never voting again, that might mean that as many as 10-15 percent of unregistered voters could be convinced to register in 2020.

There’s at least one artificial intelligence project devoted to finding and registering unregistered voters: the Electronic Registration Information Center has, in the past ten years of its existence, identified upwards of 26 million people eligible to vote but unregistered. They have done so while also “cleaning up voter rolls” (removing dead, relocated, or other no-longer-eligible voters) without “purging” the rolls of people who actually are eligible to vote.  The success of ERIC demonstrates that we can increase voter access and improve voter roll accuracy at the same time, a great thing given recent battles over other purges that have been open to charges of partisanship.

But campaigns may not be able to access such data, and there are less resource-intensive ways to figure out who’s not on the voter rolls who should be. As Adriel Hampton pointed out a couple of years ago, one way to reach the unreachable voter is by comparing official state lists of registered voters with Accurate Append’s lists of adult consumers. Services like ours can provide phone numbers and, in some cases, email addresses for unregistered voters based on those comparisons.

Consumer data is a promising comparative tool to find additional people living in an area who are not registered. It’s compiled from a variety of sources and can paint a more comprehensive picture of people’s behavior and preferences than voter data can–and it’s often grouped according to traits that campaign analysts will find useful, such as individual-level weath scores and green scores.

The Pew Report concludes on a hopeful but cautionary note:

more than 40 percent of the unregistered cared who would win the presidency in 2016, and some indicated that they could be motivated to register in the future, though many also feel that the voting process does not affect the way governing decisions are made.

So some of this is about finding the right tech and datasets, but a lot of it is about making the political process itself more engaging. Who wouldn’t support that?