What is A/B Testing in Marketing?

By on September 16, 2014

ab-testingIn marketing, A/B testing (or A/B split testing) is a term used for experimenting on two versions (A and B) of a design element (of ad, web page, email, sales copy or any other marketing element that needs to be tested) against a defined metric. Both the versions are implemented simultaneously and the one producing better result is adopted for further implementation.

For example, if an ad headline is to be tested for its effectiveness in generating more leads (well defined metric), then two versions of the same ad, with two different headlines (A and B) are released in the same newspaper in the same space on the same day. Half of the copies of the newspaper carry one version, the other half carry the other version.

What can be tested with A/B Testing?

Just about anything. It eventually depends on your goal. If your goal is to increase leads from a signup form on your website, then you will test the various design elements of the signup form page – length and persuasiveness of the copy, images used, number of fields in the form, credibility proofs, etc. All of these elements can be tested one by one with A/B testing by creating two versions of the page with only one element changed or removed. By doing this exercise you can determine whether the length of the copy is to be shortened, whether more images to be used on the page, whether the form needs to be made simpler with fewer fields or whether you should add certificates on the page as a “proof” as a proof of your authenticity.

Typically A/B Testing is done for

  • Advertising copy or headline
  • Design and layout of a website
  • Signup form page content on a website
  • Email design
  • Product pricing and promotional offers
  • Ad landing page content

Factors for Successful A/B Testing

  1. Control other factors carefully – For a successful A/B testing it is imperative that except for the elements to be tested all other factors need to be constant as far as is possible. If you insert one version of the ad on one day in a newspaper and another version on another day, it is possible that the variations in leads maybe due to other factors beyond the element being tested.
  2. Have a well defined and statistically measurable goal – You can really test how good a website looks through A/B testing because aesthetic appeal is subjective and cannot be measured statistically.
  3. Don’t depend on only one A/B test – One test maybe a fluke. More than one test giving the same result will be conclusive evidence.

About Vicky