Tuesday, April 22, 2014

Three Tips for A/B Testing With Toll-Free Numbers

Young marketer testing out a theorySome people say there’s an art to successful marketing, but there’s also a good bit of science that plays into it. Let’s say you're a dealership with a theory that car buyers are currently more concerned with a vehicle’s safety features than gas economy.

Like any good scientist, you’ll want to test your theories about your consumers’ behavior by setting up an A/B test. For instance, you could create two ads, one focused on vehicle safety and the other praising the car’s efficient gas economy.

The goal of your test is… whatever you want it to be. It could measuring everything from web page impressions to phone calls. If you’re in the mood to discover which ads give you the best phone conversions, assign a toll-free number (TFN) to each one. When the calls pour in, you’ll be able to measure your response rates easily and accurately.

Here are a few more things to keep in mind when A/B testing with TFNs.

An A/B Test Without a Goal Isn’t a Test


Always set your goals first. That way, you can create a test that accurately measures them. After setting your goals, stick to them! Many marketers fall into rabbit holes on the way to finding the results they’re looking for. Stay true to your initial experiment and make notes of the other ones you want to perform in the future.

Don’t Be Surprised If Your Test Affects Other Metrics


While you may be focused on measuring phone conversions for your experiment, you don’t want to disregard other metrics that are affected because of your changes (especially if they’re negative). For example, your ads may generate more quality phone leads but they may not be from the demographic you’re targeting. Take all of these factors into consideration when running your tests.

Sometimes a Third or Even Fourth Try Is the Charm


Just because you don’t see the results you’re looking for with your first A/B test doesn’t mean that A/B testing is just not for you. Instead of throwing in the towel, consider testing variants you didn’t consider during your first round. 

Have you had any success with A/B testing? Share them in the comments below so we can all learn from them!