In crafting their direct contacting strategies, campaigns generally take a two-pronged approach: they seek to mobilize their base and to persuade voters who are persuadable. But there is an interesting asymmetry in a campaign's ability to perceive these two target audiences. Whereas some public records in some states reveal to campaigns which voters are likely to be base supporters, there is essentially no data available anywhere in public records, or in commercial records, or in party records, that reveal to campaigns which voters are persuadable. Persuadability is a disposition that scholars of political psychology have explored by analyzing public opinion surveys. These scholars focus on voters' level of political awareness, knowledge of political information, cross-cutting positions on issues, and value systems as dispositions that indicate one's latent persuadability or fickleness of opinion. The trouble is that no information that campaigns have about voters allows them to connect with voters who are persuadable as measured in these ways.

The asymmetry between data resources predictive of voters' partisanship and predictive of voters' persuadability squares with research findings about the effectiveness of direct voter contact. Scholars and political practitioners who have measured the effectiveness of voter contact have established a clear set of best practices for get-out-the-vote (GOTV) efforts (Green and Gerber, 2008). Collectively, ground campaign experts know a lot about how to increase turnout among known supporters. In sharp contrast, there are no equivalent best practices for persuading voters. Experimental studies on persuasion show these efforts are often ineffective and sometimes even counterproductive (Bailey, Hopkins, and Rogers, 2013). In part this is because it is easier to measure the effectiveness of GOTV efforts than persuasion efforts: turnout is a public record; vote choice is not. Here, I point to a different reason why persuasion efforts are difficult. In campaign databases, there is no identifiable group of voters who are predictably and consistently persuadable. In other words, persuadable perceived voters are a very noisy approximation of actual persuadable voters.