Data Mining For Dummies
Book image
Explore Book Buy On Amazon

Most political campaigns depend on consultants to provide voter research, or else get by with very informal assessments of voter attitudes and interest in voting for a particular candidate (or voting at all). But in recent years, certain political campaigns, including both candidate and issue campaigns, have begun to use microtargeting, organized programs of survey research and message testing, to develop and deliver personalized campaign messaging tailored to individual voters.

Treating voters as individuals

Think of the difference between shopping at a mall and shopping on your favorite online store. At the mall, everyone sees the same signs, the same flyers, the same items on display. The shopper has access to everything that’s available but has to make the effort to discover the most appropriate items.

Your favorite online store doesn’t show the same ads and products to everyone. It uses your past history to tailor the presentation. In the online store, you see ads that were selected specifically for you based on factors such as the items you’ve purchased before, the products you’ve looked at, and the products purchased by others whose buying or browsing history resembles yours.

And that online store tests every element of the presentation (offer, text, images, layout, and more) to find out what works best.

A typical political campaign might use a program of political polling to identify key messages for voters as a whole, or large segments of voters, such as women, seniors, or youth. Microtargeting examines each voter as an individual and, like the online store, uses information about individuals to personalize the campaign.

Consider that two candidates, Fred Mertz and Lucy McGillicuddy, are running for office. Fred will use traditional campaign techniques. Lucy’s campaign will use microtargeting. How will the two campaigns differ?

Both campaigns will make use of publicly available voter records. These records provide each candidate with a list of registered voters, addresses, and voting history. The records don’t reveal how individuals vote! The ballot itself is always secret. But they do tell if and when someone has voted, and may include details such as party affiliation (some regions require this information for voting in primary elections).

Even Fred, the traditionalist, understands that people are not all identical. But he has little or no information about the attitudes of individuals. So, he’ll mail the same brochures to everyone in the district and use the same few messages in all his advertising.

He’ll get out to shake hands in every neighborhood. Although he understands that people in different neighborhoods have differing concerns, he’ll have only his intuition to guide him in speaking with individuals. At best, he’ll be making educated guesses about what to say to each person.

Enhancing voter data

How will Lucy’s microtargeted campaign be different? For her, the public voter records are only the heart of the data resources. Lucy’s voter database will include lots of information that Fred’s does not, such as

  • Demographics

  • Occupation

  • Political and charitable contribution history

  • Memberships

  • Home, auto, and boat ownership status

  • Permits and licenses

  • Magazine subscriptions

  • Political volunteer history and other indicators of political views

How does Lucy get all this information? Her political party, private sources, and her own team enhance the voter database with additional information about each individual voter. Some of this information is available through public records and some can be purchased from commercial data vendors, but the most valuable information for a political campaign comes from one-on-one contact with prospective voters.

Gaining an information advantage

Assembling Lucy’s voter database involves a lot of work! Even a candidate who has the money and know-how to obtain data from disparate sources and match it to individual voters would still need a lot of patience and labor to integrate data sources. Few campaigns have such resources.

Fortunately for Lucy, her party has already developed an enhanced voter database that she can use as a starting point. So she starts her campaign with a significant information advantage over Fred. But plenty of data gathering is still ahead for the microtargeted campaign.

What about Fred? Why isn’t his party providing data to help his campaign? Major political parties in the United States, Canada, and the United Kingdom now all have voter databases for their own candidates, so many candidates could launch campaigns with data resources similar to Lucy’s. If Fred isn’t doing this, perhaps he’s unaware of what his party offers, doesn’t know how to use it, or just doesn’t appreciate the value of the data.

Developing your own test data

Although Lucy now knows a lot about the voters, she doesn’t yet have the information that she needs to tailor messages to individual voters. To get that, Lucy and her campaign team must conduct an ongoing program of developing and testing specific messages.

First, her team will identify some major voter segments using the enhanced database. Perhaps they have data from some preliminary surveys that they will use to divide voters into three groups to start: firm Fred Mertz supporters, firm Lucy McGillicuddy supporters, and undecided (or persuadable) voters. Within the undecided voters, they might next choose a narrower segment, such as Latina working mothers.

They could then brainstorm about the issues and messages that might appeal to Latina working mothers. The marketing team would develop some sample scripts, each focusing on one specific issue and message. The messages must be consistent with the candidate’s position on the issue, but plenty of options will exist for investigating which issues to highlight and which messages are most persuasive.

The only way to know what works is to test. In a typical test scenario, volunteers would be given lists of voters to call and alternate scripts to use, such as one focusing on public schools and another on health clinics. The volunteers would read from the script and also ask questions about the voter’s likelihood to vote for Lucy. At the end of the test, Lucy’s campaign will have new data, the survey responses gathered during these test calls.

Lucy’s campaign now has unique survey data. Using this data, Lucy’s campaign now discovers which of the test messages was most persuasive to a specific group of voters.

The survey may uncover details that were not specifically part of the test. Voters may reveal something unexpected in their comments. Some of this group might mention that they are not so concerned about public schools because they send their children to parochial schools.

That information helps the candidate to understand why certain messages work better than others. It also hints at the opportunity to get even deeper in understanding individuals. The next survey might compare Latina working mothers whose children go to public schools with Latina working mothers whose children go to parochial or other private schools.

Taking discoveries on the campaign trail

Now that Lucy knows something about a message that appeals to Latina working mothers, she and her volunteers are going to use that information in everything from Lucy’s speeches to flyers to talking points for volunteers canvassing neighborhoods.

Lucy’s team will conduct surveys like this, on the phone, by email, and face to face, throughout the campaign. Her database will be augmented on an ongoing basis with new voter data and new tests. As the depth of Lucy’s information grows, she and her volunteers will become better equipped with each passing day to present individual voters with messages that are relevant and appealing to that specific voter.

About This Article

This article is from the book:

About the book author:

Meta S. Brown helps organizations use practical data analysis to solve everyday business problems. A hands-on data miner who has tackled projects with up to $900 million at stake, she is a recognized expert in cutting-edge business analytics.

This article can be found in the category: