A/B Testing

More than a specific experiment, A/B testing is a way of working that encourages you to test 2 variants A) and B) of a certain element.

By running an a/b test you can discover which performs the best in maximizing the desired outcome. It is important to decide upfront what this desired outcome should be. This could be driving (potential) customers to your website, a visiting time on your website, etc.

There are many elements that can be a/b tested: 2 different subject lines of a newsletter campaign for example, 2 versions of an ad campaign, 2 pricing schemes of a new service, 2 colors schemes for a landing page, and so forth.

You should go through the following steps to run an a/b test:

Research: start with researching what exactly you want to a/b test. Identify ‘problem’ areas that you want to optimize.
Observe and Formulate Hypothesis: create data-backed hypotheses aimed at increasing conversions.
Create Variations: create a variation based on your hypothesis and A/B test it against the existing version (control). A variation is another version of your current version with changes that you want to test. You can test multiple variations against the control to see which one works best.
Run Test: there are several tools you can use to run an a/b test (some examples can be found below).
Result Analysis: once your test concludes, analyze the test results of the metrics defined as your desired outcome. If the test succeeds, deploy the winning variation. If the test remains inconclusive, draw insights from it, and implement these in your subsequent tests.

Tools A/B Testing

Convertize

VWO

Hubspot

Video A/B Testing

Any Questions?