The Importance of Testing, Obama Style

imagesThere’s a recent article about the dramatic difference in results between the best and worst performers tested in President Obama’s fundraising emails during the presidential campaign.

In short, there was a 600% performance difference between the best and worst subject lines used during their testing process.

While these numbers are dramatic, they are not surprising. We see test results on a daily basis and whether it’s an email subject line or an A/B test for an ad, it’s rare that we don’t see one version performing dramatically better than the version(s) being tested against.

The takeaway? If you’re not testing, you’re losing money.

What are some of the most accessible testing strategies across a variety of mediums?

Subject line testing in email (like the Obama effort)


Even if you don’t have a sophisticated email provider (to randomize and split your lists into testable segments), you can randomize your list in Excel and then break it into an appropriate number of groups to test different subject lines to. When testing subject lines, experiment with different tactics like:

  • including the recipient name (if possible).
  • asking a question.
  • using different writing styles.
  • highlighting different benefits (of what’s inside).

Ad testing (in Google Adwords)


It’s amazing how few people test their ads. If you’re running ads based on what you “think” people will respond to, you are not getting the best performance you could. Plus, it literally takes only seconds to add additional ads to your ad groups (and Google automatically takes care of the rest).

Adwords testing has other huge benefits to companies launching new products or services. If you want insights on which user benefits are the MOST important to your audience, simply run a series of different ads, each highlighting a different benefit. You’ll be able to tell pretty quickly on the basis of your clickthrough rates which of the benefits resonate most with your audience! You can then take this knowledge and apply it to your product/service pages with some confidence that you’ve identified the right benefits.

Web page testing


Ever argue with your colleagues about what should go on a new product page? Argue no more – test instead!

With Google Content Experiments (in Google Analytics) you can create multiple different versions of a page, have them randomly served to visitors, and identify the top performing version. While it’s not necessarily “easy” to set up, it’s well worth the hour it takes to learn to configure “Goals” and set up your Content Experiment.

Questions about testing? Drop us a line.

Similar Posts