A-B testing is a great way to compare online marketing strategies or tools to see which one is working better. It’s a cost effective way of pitting everything from web pages to email campaigns against each other so you get the best ROI. Using this method, you’ll get the numbers and statistics to see what needs to be changed out, tweaked or left alone.
Here’s How A-B Testing Works
Let’s say you’re putting together a website for your small widget business. You’ve sourced a few different designers and got two excellent mock-ups. Your torn between the two but need to make a choice.
A-B testing allows you to keep your emotions in check and use empirical data to make a decision. You’ll need to start by splitting the website traffic between the two candidates.
Then, once the data starts flowing in, you can start to see which website designer’s work is performing best. With this example, you’ll also need to herd the numbers toward certain corrals that matter to you. For example, you might use metrics like the conversion rate and the bounce rate of both designers before making a final decision.
The metrics that you use often depend on what you’re testing. However, there are a few common ones.
Bounce Rate
In a nutshell, making a good first impression is what this metric is all about. That’s why you need to include it. If people are looking at your landing page(s) and leaving right away, it’s a big red flag you need to look at.
Exit Rates
These tell you the story of the visitors that get past the landing page but still decide to leave. If visitors are dropping off at a certain page, you know where to start work.
Engagement Metrics
These are averages that help to bring everything into focus. Taking a look at the averages can show you which pages need to be tweaked.
Like anything else you use for small business marketing, there are some do’s and don’ts when it comes to A-B testing.
A Few A-B Testing Don’ts
Don’t test one item and then the other. For example, if you’ve got two email campaigns to choose from, testing one in September and one in October will skewer the results. Running both at the same time ensures the subjects or traffic is consistent.
Don’t be in a rush. You’ll be getting lots of information once you start one of these tests. The trick is not to end the experiment too early. Using only a few visitors over a short period of time wont give you enough data to make the right choices. Here’s a calculator that will help you decide how long the test should be.
A Few A-B Testing Dos
There are some best practices you should follow.
Be consistent. If you’re testing a Call To Action across several pages, it should designed the same across them all. Changing the design from page to page will skewer the data.
De several tests. The chances are you’ll make a few mistakes on your first A-B test. To get a template you can use, you’ll need to work on refining the techniques you use. For example, you can test several variants at once or even design your own approach. Get a few good ideas here.
AB Test Photo via Shutterstock