Here’s a statement not even a candidate can debate: political marketing is tricky. To know for certain what is and what is not working in your political marketing plan you need to test your methods and track those results. So, let’s examine the basics of good marketing and rewrite the rules as they would need to be applied for political initiatives.
Here’s a sentence you need to memorize: You can’t test a test. By this I mean you need to have one proven factor (a constant) and test one new factor (a variable.) Testing in this manner enables you to quickly and efficiently identify elements that produce improvements.
To start, let’s think about testing creative. You could also test wording, solicitation amounts or methods, whether a policy email elicits more response than straight donor request, etc. But whatever it is that you’re testing, you are only testing only one element—one test, one variant. Another thing to remember: don’t just broadcast out one new-variant email. You’ll never know if it was the variant or the list that did or did not work.
What’s the best way to structure a test?
The first step is to pick an audience that you know has worked for you before (a constant) and then do what’s called an “A/B Split.” Half of your constant will get the standard version of your messaging and the other half will get the same message with your new testing variant added in. Make sure that every single thing about the email is exactly the same—except the variant. If you start changing wording or other factors you’ll simply never know what worked and what didn’t.
So you send your A/B Split email out to a constant audience doing a random “n-th” split of the list, which just means you set a structure of how to pull the random list—separate every 5th name, for example.
You also need to come up with something that will evoke a response. If your traditional email message is a donor solicitation or a statement about your policy then you need to come up with one constant “giveaway” or something similar aimed at evoking more of a response. If people don’t have a way to take action you cannot track your test. And, in order for it to be a valid test, both groups in you’re A/B Split have to take action.
You could give away a t-shirt, hat, or a research report. You could offer tickets to a meeting or a breakfast. Basically, there has to be something that someone will respond to so that you can determine what worked in your email messaging and what did not.
Some potential tests:
- Test a donation solicitation with and without further statements on stands, experience, background, policy, etc. Remember to not flood the email with text—choose one item to test against the donation only email to see what pulled in the additional response.
- Test your response mechanism. How do people actually respond? Do they click? Do they call? Are they signing up with a form?
- Test the creative look of your emails. You can test everything from the colors to the fonts to the pictures. These all evoke some response from an audience. Not sure if you should use a photograph of your candidate alone, with the family, in the office, or at home? The right test could help you find out what your constituents want to see.
- Test subject lines, test “from” lines, and test timing on your emails—when you press send can have a major impact on response.
Testing teaches you about your constituent audience. But before you rely too heavily on marketing tests to learn about your voters, flip that equation. Take the time to learn about what motivates your audience and then structure tests to fine tune your approach. In this way you will truly zero in on the most effective political marketing techniques for your campaign.
Carol Lustig is CMO at Media One, a data provider with a specialty in U.S. voter registration and constituent data.