There are many ways to win an election, but two are tried and true:
|Persuasion||Get Out The Vote|
|Identify the people who will vote on their own, then persuade them to choose you.||Identify the people likely to favor you, then convince them to cast ballots.|
Most campaigns start with persuasion, but as election day approaches, we switch to get out the vote. For many years, I've been writing software to manage these two tasks.
For me, each election is an opportunity to test my software and the theories behind it. Between campaigns, I scramble to write code for everything I learned in the previous cycle. But I never seem to catch up!
When you become a candidate in most states, you can purchase registration data for everyone eligible to vote in your race. Prior to writing software to handle this data, our volunteers had a lot of enthusiasm and worked like crazy, but we were often beaten by candidates who put in less effort, but worked smarter.
Here is part of the demographic analysis that Village Green generated for the November 2012 election:
Likely voters, 3 out of 5
Perfect voters, 5 out of 5
Both of these charts describe San Francisco voters. On the left are those who voted in 3 of the last 5 elections. On the right are people who voted in all 5, a category that I call ‘perfect voters’.
If you are running for a lower level office, perfect voters are crucial. While Facebook might be a good way to reach the likely voters in the left hand chart, you'll need other methods to reach the ‘perfects’ on the right.
Village Green can do more advanced analysis as well. Below is a simplified network of voters in San Francisco's District 9:
Each circle represents a specific voter. Political party determines their color. The size of a voter's circle depends on that person's potential, a numerical measurement of the help they've given to campaigns in the past. Full Village maps are interactive and contain thousands of voters.
Voter rolls are divided into districts and precincts, and election results are reported that way too. Below, Village Green has shaded a map of San Francisco according to the results of a ballot proposition. Areas where voters chose 'yes' are green. Red is where they voted 'no'.
We can use the results of past elections to predict voting in the future. We can do the same with geographical census data. By cross referencing many different databases, we can gather quite a lot of information about each voter, and then Village Green can put it all together in a mathematical model.
The end results are lists of people to contact at each stage of the campaign.
Most of the work in a field campaign is knocking on doors and calling voters on the phone. We often assign precinct walking to the younger volunteers, and older folks make the calls. Since some older people are uncomfortable working with computers, I print paper lists.
Village Green can generate a variety of different layouts, but here's one example:
|and many others…|
Each sheet contains several households that look like that. As a caller, you work down each page drawing circles on the paper according to the results of each call. There's also empty space to make notes. Later, we record these hand-written marks in the database using my bar code tool, Canvass.
Very soon, I hope to supplement the paper driven systems with web, phone and tablet apps. But it may be a while before paper is completely replaced.
On the right is a partial log of activity during a recent campaign. We often hear that the politician with the most money is the one that wins. But I've found that a team of volunteers like this, willing to make hundreds of calls, is fully capable of victory.