Four A/B Testing Mistakes to Avoid

Should your “Subscribe Here” button be red or green? Should your headline be short and punchy, or more descriptive? Should your landing page have only one form field, or will just as many people convert even if they need to fill in three fields?

Because every business and every group of consumers is different, we can’t tell you the specific answers. But there is a way to find out: A/B testing.

By using today’s closed-loop marketing technology, which makes tracking prospect behavior easy, you can test your emails, landing pages, calls-to-action and more to find out what variables truly influence click-through and conversion rates. But you need to keep a few best practices in mind if you want truly meaningful results that will help you optimize your marketing efforts. Don’t make these mistakes:

1.  Conducting multiple tests at once.

Because your boss is breathing down your neck to increase lead generation numbers, you feel pressured to improve email click-through rates AND improve landing page conversions ASAP. In order to do so, you decide to A/B test both your next outgoing email campaign and the landing page it points to. So you try two different email subject lines and randomly show landing pages with only on form field or with three. Leads increase, but do you really know why?

Because you tested multiple marketing elements at once, it’s going to be tougher to tease out whether the increase was due to fewer form fields on the landing page or a stronger email subject line.

Limit your tests to one element at a time, especially if you are new to deciphering closed-loop marketing and A/B testing results, to give you a clearer picture of what changes will improve your lead generation efforts.

2.  Testing multiple variables at once.

In your marketing meeting, one person suggested that marketing emails should include an image, which you haven’t used in the past, while another insists that the key to a higher click-through rate is including multiple links to the targeted landing page. Thinking that both are good ideas, but need to be tested, you send two versions of an email—one with no image and a single link to the landing page, the other with an eye-catching photo and multiple links to the landing page.

Again, the new version improves your CTR, but you don’t know which change made the difference. Isolate a single variable and test it for a sufficient length of time to get meaningful results.

3.  Overlooking small changes.

If every test you run is based on major differences—HTML emails vs. text-only, or landing pages with video vs. landing pages without—you may be missing out on little changes that can make a big difference.

Sometimes, something as simple as the color of button can make a big impact, as HubSpot and Performable found in an A/B test on Performable’s home page. They tested a simple call-to-action button with the words “Get Started Now!” using versions in red and green. What they found, much to everyone’s surprise, was that the red button outperformed the green button by 21 percent.

The only way to know that is to test even the smallest of variables.

4.  Being afraid to test something big.

On the other hand, sometimes you need to test big changes, such as overhauling an entire landing page format or greatly increasing the perceived value of the offer.

In its early days, HubSpot offered its visitors a seven-day free trial of the software. But in 2010, they decided to see what would happen if they tested a 30-day trial instead. Their A/B test showed that majorly changing their free trial offer more than doubled their visitor-to-trial conversion rate.

How many potential leads would they have missed out on had they not tested such a big change?

Make sure you are avoiding these A/B testing mistakes.  Or are you not testing your marketing elements at all? The solutions, along with step-by-step instructions for setting up A/B testing of emails, landing pages, calls-to-action and more, are in our ebook Introduction to A/B Testing for Marketing Optimization. We’d also be glad to show you around the HubSpot software to see how simple it can be to run your A/B tests. Request a demo today.