Four Hot Tips for A/B Testing

Something that many of my clients neglected before I came along was testing. They didn’t have any idea how to accomplish it, or even that it was important. Yet, when I explain the benefits of testing to them, they’re all about learning more and fast!


Because testing = more money. That’s it! You should test everything about your pages — headlines, bullet points, images, calls to action, etc. You will find out what works best and in time, make better conversions because of it.

A/B testing

Though I have written about this topic before, my friend, Ruben Corbo, has some new  advice about the simplest type of testing — A/B Split testing and he’s written a great post for you to consider and put into practice. Here goes:

The 4 Most Effective Tips for A/B Split Testing

As an entrepreneur, you probably wonder how to expand your market share and outmatch the competition by providing products and services that clients need and relish—while making money in the process. Right?

“A/B testing, a byproduct of advanced statistics, helps you forge an effective marketing and product differentiation strategy,” says Edward G., a marketing specialist who for several years has advised clients on everything from A/B testing to multivariate testing and behavioral targeting.

Don’t worry, though; you don’t need a PhD. in math to understand and use A/B testing. By applying a few key tips, you can implement a successful, fact-based A/B testing routine.

1. Understand A/B Testing

A/B testing involves changing only ONE element of your page, while leaving all else the same, in a closely monitored experiment. For example, say you are getting ready for the Thanksgiving sales season, and you want to add a turkey image to your call to action button, hoping that its familiarity and charm will encourage people to buy.

Create two versions of the page, with Version A being the “control version”—the one that does not change—and Version B being the one in which you modify only that one aspect (add the turkey) and then, run the pages side by side in a kind of rotater. Visitor A sees the control page, while visitor B sees the new and improved turkey version.

Then, it’s important to keep watch on your stats. How many people clicked through to buy in a certain period of time? More clicks on the new page means that the turkey was successful, so that page becomes the control and you can run that page against a new page with ONE different element  against it.  

2. Learn What A/B Testing Entails

A/B testing—especially for Internet marketing, which is what we are focusing on here—entails two versions of your website, hence the “A or B” concept. Next, you should launch both portals simultaneously or at approximately the same time interval.  People buy more or less at different times of day and if you run your pages at different times of day, along with the change you made to the page, you won’t know what worked to increase your conversion. So, launch both pages at the same time, if at all possible.

Keep everything intact in both versions, except one element. Generally speaking, you don’t have to create two websites from scratch. Just copy the page you want to test and then, change ONE thing about it — the headline, the bullets, the call to action, the background color, etc. Choose ONE element to test at a time, or again, you won’t know which element made a difference in your rate of conversion.

3. Identify Key A/B Testing Criteria

When it comes to A/B testing, criteria are as varied as colors in a rainbow. It all comes down to what you want to evaluate, so the sky is the limit with respect to choices. That said, Edward G., our A/B testing expert, says that a few elements are essential to draw up and implement an effective A/B testing campaign. These elements, aside from those mentioned above, can also include product or service descriptions, pricing, special discounts and other promotions, and content length, but really, anything that can be changed is an element for testing consideration.  

4. Familiarize Yourself with the A/B Testing Process

As explained earlier, A/B testing is a straightforward process, but you still need to follow specific procedures to make it a success and you need enough traffic to the page to have statistical significance. If you’re not terribly skilled technically, there are online tools and companies that are available to help you.  Having a company test for you will be well worth the expense in increased revenue that will come from it.

Here’s a quick review:

  • Review your existing content and determine what to alter, improve, or remove.
  • Come up with a list of elements that you want to test over time. You can categorize the elements in the 4 marketing-mix clusters: product, price, promotion, and point of sale.
  • Figure out key testing elements (Make the most important changes first.)
  • Test both versions of a page simultaneously or enlist a professional to help jump-start the process.
  • Assess the results. Then, keep or discard the tested elements. If you keep the tested element that changed page becomes your new “control.”
  • Begin the process all over again with the second testing element on your list


A/B testing can be simple or as difficult as you want it to be, but you will be successful,  if you understand the process and familiarize yourself with its main stages. You can also hire a testing professional or company to help you in the process. This is a painless way to handle testing, and though it will cost you, the gain in improved conversions will be well worth the expense.

Testing isn’t a quick fix for a non-converting site, but ALWAYS be Testing. Even if your sales page is converting, why not allow it to convert better? It’s all about making money and testing is one surefire path you should take.

Ruben Corbo is a freelance writer and writes very frequently about topics in the Internet marketing arena. Ruben specializes in writing about traffic to online sales conversion tools like A/B Testing, Multivariate Testing, Product Recommendation Engines, and more. To learn more about these conversion tools, go to