A or B?
A/B Testing AKA Split Testing is about arranging a battle between two versions of your landing page. Then you make your bets (highly optional) and see which one is doing the best job leading your users to the target you put have chosen like subscribing for example. You are free to test two completely different versions of the page or the ones with slight differences. It does not matter.
A/B testing may be a great mean of improving your site-visitors communications as well as mobile application development. It also is great with backing up changes and important design-related decisions with actual metrics. And with the tools that are now available Split Testing does not even require too much of a technical basis from a tester.
When to go for A/B?
You will require visitors for your Split Testing sessions. You will require an appropriate amount of conversions. These numbers will rely on your site’s personal number of visitors. In order for A/B to mean something you will need the least of 1.000 users per every option as well as 150 conversions for every single variant. These numbers are achieved within hours with certain sites and with month for others.
Remember that your particular business may be happy with the total of 100 visitors per month. And that is great for you. Just don’t engage into split testing if that is your case.
Start small. If it is your very first Split Test make it as simple and painless as possible. Here are some ideas to begin with.
- H1 Versus H3 headings
- Call-to-action button stating ‘subscribe now’ versus ‘sign in’
- Short versus long page (should you desire to hide long sections)
Never forget that you can run several scenarios at the very same time without any damage whatsoever. Don’t get stuck with only one change at a time. Yet don’t get carried away or confused at the start. Once the winner is determined make the changes permanent with updating the underlying core. And now you have something your users have personally chosen.
Tool to use
- VWO (Visual Website Optimizer)
- Google Analytics Content Experiments
- Campaign monitor (this one is for e-mail campaigns)