How to Get Started With A/B Testing

A well-known marketing maxim is "test and refine, test and refine." Without testing, you can't know what's working and what is not. Fortunately, the testing process is very transparent in digital marketing because you can see all sorts of campaign-related data in near real-time and tell what's working at a glance.

One of the best ways to boost the effectiveness of your marketing campaigns and improve conversion rates? A/B testing , also known as "split testing".

How can you as a small business owner run such tests without a team of specialists? We'll break it down in this post and explain how to do A/B testing for email, landing pages, ads, and other marketing efforts.

A/B Testing Defined

A/B tests involve making subtle changes to key components of a similar message.

Let's say you are planning an email campaign to promote a sale. You create two versions of the same message and send one to a segment of your mailing list and the other to a different segment, and then wait to see which gets the most response.

Many email marketing platforms automate this process:

1. Choose a percentage of the list that should receive each version of the message (20 percent is a useful sample size).

2. Send each version in equal numbers

3. The platform calculates the results and then sends the version that performed better (the "winner") to everyone on your list who didn't receive one of the test emails.

You can use a similar process for web pages, landing pages, digital ads and other marketing collateral. You show two variants to similar visitors at the same time — the one that converts better wins.

A/B Testing Best Practices

Here are four ways to make sure you get the best results from A/B tests:

1. Know the Purpose for Testing

It's possible to test virtually every element of a marketing message. Before deciding on a variable, however, know the purpose for testing. Is it to improve email open rates, get more clicks, increase conversions or sell more products?

2. Stick to a Single Variable

Don't be tempted to make too many variations to your message as that only complicates matters. Instead, a single change makes the differentiating factor much easier to pinpoint.

3. Create a 'Control' and 'Challenger'

A/B testing is not unlike the science experiments you ran in high school. Once you know the purpose of testing and have determined which variable to test, create a "control" version and a "challenger" — the page you will compare against the control.

For instance, if you wonder whether including a customer testimonial on a landing page could increase conversions, create the control page without the testimonial but include it on the challenger page.

4. Use an A/B Testing Tool

Chances are, the email and marketing automation software you currently use supports A/B testing.

In the event yours does not, either consider switching to one that does or use a third-party platform designed for such use. VWO, Optimizely, and Instapage are popular options but come with a price tag. Google also has a testing tool called Optimize, which is free.

Elements to Test

With the myriad of elements to test, which will likely have the greatest effect? Consider these seven:

1. Call to Action

Your call to action (CTA) ultimately drives conversions. A/B testing of your CTA can often lead to higher rates.

These tips should help:

  • Reword your CTA to make it more direct.
  • Vary the specific language — do your visitors respond better to "Add to Cart" or "Buy Now"?
  • Play with your CTA button — often, a larger button works better than a smaller one.
  • Move your CTA higher on the page (if it's not already above the fold).
  • Repeat your CTA multiple times.

2. Email Subject or Headline

Your email subject line or headline (for a web page, landing page or ad) must grab the attention and entice visitors to read further. Test one that is forceful, direct or problem-focused against one that's more of a soft sell.

3. Page Copy

The page copy should deliver on the promise contained in the subject line or headline and convince the visitor to act. Examine the text carefully with a critical eye. Does it explain how your offer will fill a need or solve a problem? If not, revise it, and then test the revision.

4. Images

The images you use can evoke an emotional response that leads to the sale or lead. A good way to test images is by using a photo of people in one version and a product in another. You can also try using an image versus a video to see which performs best.

5. Colors

Colors can also affect how a person responds. Try testing stronger colors such as red or orange against softer, pastel tones.

6. Forms

People are often reluctant to give out personal information, so test the use of a shorter form (e.g., name and email address) against one that asks for more.

7. Mobile-friendliness

We live in a mobile world, which means your marketing collateral must be mobile-friendly. Experiment with the length of mobile pages, adjust display options, and alter the navigation.


You can't know what part of your marketing is working unless you test.

A/B testing can be a boon for any small business owner willing to spend time learning the ropes and experimenting with the many options.

Make sure to know the purpose for testing, start with a single variation, and use the control versus challenger approach. Thankfully, software platforms are available to make the process easier and less time-consuming.

See all
Posted on Date:
Wednesday, December 12, 2018