Sharpen Your Marketing Message With A/B Testing

Babak Ghoreyshi

The world may be changing, but the digital world is changing even faster. In today’s “wild wild world,” if you want only to survive, you will lose. Aim harder to win the challenge.Businesspeople Looking at Graph in Meeting

In this fast-moving era, marketers are always striving to offer the best content through the most relevant channels at the right time to the customers and prospects. It is difficult — maybe impossible — to say which digital strategies and tactics are best because companies have different segmentation models, different target groups and different values to provide.

There are always multiple solutions, and the only way to know which ones work — and which don’t — is to test them. Fortunately, A/B testing enables us to find the right options. However, just doing the A/B testing isn’t enough. You need to know how, when, and why to use it.

In their book, “Emarketing Excellence: Planning and Optimizing your Digital Marketing,“ Dave Chaffey and PR Smith define A/B testing as “testing two different versions of a page or a page element, such as a header image or a button. The goal is to increase page or site effectiveness against key performance indicators.”

A/B testing can be extended to A/B/C testing, which includes the neutral model or existing status of the page and enables the marketer to evaluate two options against a neutral situation. There is also multi-variable testing that evaluates a combination of different elements. In the multi-variable model, if the variables exceed a certain number it will be difficult to interpret the data without using a sophisticated tool and evaluation model.

Three words best describe the concept of A/B testing: Test, learn, and refine. This test enables marketers to track, monitor, and select the best methods to communicate the message. There are different attributes to consider when changing or optimizing each tactic that can increase conversion rate fundamentally. Starting with forms in lead-generation campaigns, we should consider and test these five dimensions:

  1. Optimize form length
  2. Mind the colors
  3. Tweak form copy
  4. Improve user interaction elements
  5. Show social proof, testimonials, and endorsements

For email marketing, newsletters, and communication with customers, marketers should consider the following 15 elements to ensure that they are using the best option:

  1. CTA
  2. Headline (this can be your last chance to nail it and grab attention, so don’t miss it!)
  3. Body text
  4. Subject line
  5. Having testimonials
  6. Personalization
  7. Design (multi column, pic, plain…)
  9. From Name
  10. Pre-header text
  11. Soft sell (benefits) versus hard sell (buy now)
  12. Font (type, size, and color)
  13. Responsive versus non-responsive design
  14. Send time (day and time of the day)
  15. Offer type

Don’t forget to test your landing pages! Always start with:

  1. The main headline
  2. The call to action (CTA)
  3. Product or service shot
  4. Button design
  5. Color of the button, scheme, and frame fonts
  6. Form length
  7. Long copy vs. short copy

“Three areas of marketing that have been transformed by digital are the speed, relevance and reach of campaigns.” -The Guardian

To stay relevant and manage the fast changes of tech trends, customer behavior, and channel usage, marketers need to make sure that each single campaign is developed with the right message for the right person at the right time and is distributed through the right channel. All elements of a successful campaign should be optimized based on the trends, needs, and personalized behavior of the audience.

There is no single rule of thumb, nor any absolute rule in the digital world. Do the A/B test to find the answer and the best solution, but don’t change too many form elements at once. Remember that reality is not always what we perceive. Test one variable at a time and go forward step by step. For a successful campaign, even minor things can be tested and can change the final results fundamentally.

One last tip: Split your sample group randomly. Without randomness, this experiment can’t help you.


About Babak Ghoreyshi

Babak Ghoreyshi is a Solution Manager at SAP.