yc-blog-blogheader-12.7.18

Digital Marketing – Boosting Conversions is as Easy as A or B

Boosting Conversions is as Easy as A or B - 12.27.18

In your estimable opinion, your landing page is pure gold. Your staff concurs – at least, publicly. But will that glorious page hold up against the scrutiny of picky visitors? Fortunately, brand marketers can capture some convincing answers with A/B testing.

A/B testing is an exceptionally reliable experimental process. Basically, it offers a single choice to one set of website visitors, and a different choice to a second. In the best tradition of outright sneakiness, neither group knows about the other. Case in point — pre-selected visitor Group A sees one landing page upon arriving; Group B sees another page.

Selecting group members is up to the testing brand, of course. Criteria to be determined. Ultimately, what’s important is the information gained – one of the two landing pages will generate a superior response rate easily measurable as conversions. And that page, all things being equal, should be chosen for permanent duty. Based on available evidence, it will produce more conversions than the alternative.

Brand marketers, bear this in mind — A/B testing needn’t be limited to landing pages, or any other type of page, for that matter. The technique is suitable for many other components – everything from taglines to CTA’s may be offered in A-B test pairings.

Is there an optimal duration for a test period? Nothing is etched in stone on that matter. Effective test periods can vary, depending on goals, complexity, customer characteristics, and other crucial variables. Test designers should consider all relevant variables to ensure the proper time frame.

Also important is quantity of test subjects. Again, no minimum sample size exists. But practically speaking, a hundred or so participants likely will fall short. In most cases, the magic number will be somewhere in the thousands – enough to cover the subject. As with similar tests, the more participants, the greater the accuracy. Something for test designers to ponder.