Don’t guess, A/B test

By 26th September 2014Digital Marketing

At LogicSpot, we love helping businesses grow. If you have not explored what A/B testing could do for your site, now is the time.

What is A/B testing?

Simply put, A/B testing is a way to test different versions of a web page to see which performs better. If you are looking to increase newsletter sign ups, or click throughs to your product pages, A/B testing could be a great way to get insight into what works well on your website.

Who is A/B testing for?

A/B testing can be a powerful tool for business owners, ecommerce managers, or any other role where you work closely with a web team. If you are creating or growing an online business you will have no doubt been in discussions about new designs and layouts, and debated which design will work better on your site.

These discussion are often times opinion based or maybe backed up with a usability article. I have worked in a variety of industries and found that agreeing on a design is not always easy. The stakeholders, designers and ecommerce managers all have their interpretations of what design or layout can work well for their user base, but these are often just presumptions. In the end, the hippo (highest paid person’s opinion) usually gets what he or she wants.

Stop guessing, start testing

A better way to make these key decisions would be to run tests and see what works well for your audience.

Let’s say you have just created a new landing page, the goal is to get users to sign up to your newsletter. With A/B testing, also known as split testing, you can produce two versions of the page, each with a different design. You can then split your visitors between each page, and track which one reaches the most goals.

a/b testingSounds great, What can I test?

You can test anything where the goal of the page is for a user to perform an action. That can be completing a form, moving through a journey (‘click next to see step 2’) or buying a product.

You can create more than one variation and you do not have to use completely separate pages. For example you could test the call to action on a ‘buy’ button.

  • Original button – ‘Buy now’
  • Variation 1 – ‘Buy it’
  • Variation 2 – ‘Complete order’

If you are stuck for ideas, here are 71 things to A/B test. (Please note, I would not recommend all of these  #29?)

How much flexibilty  do I have?

A/B testing software allows you to control all aspects of your test.

  • You can choose how long you would like to run your tests for (one week or one month)
  • You can set how many visitors see your A/B test, 50% of your traffic? 100%?
  • You can choose which visitors see your tests, for example only new visitors, or only users that are signed in?

Conclusion

A/B testing can be a great tool but it is important to run tests long enough to get statistically significant result. There are tools online that can help you with this.

I hope this quick intro has given you the principles of A/B testing. If A/B testing is something you want to try, contact info@logicspot.com

  • https://www.logicspot.com Mark Haller

    Hahaha, great article Simon … I’ve never heard of the phrase “Hippo” before! Geez, I hope I don’t start sharing that out loud with clients!