A/B Testing is something website designers should be utilising. It’s a fantastic practice which can provide you with better insight into your client’s visitors’ behaviours and how to increase and improve conversion rates for your clients.
What is A/B Testing?
The clue is in the name. A/B Testing pits two versions of a page element against each other using a metric which will show the success of each. It’s used to choose between which from A or B is the best option for your design.
For example if you have two designs and you can’t work out which is the best for your client then you can use A/B Testing. What you’d do is split the incoming traffic on the site equally towards one version and after a pre-set time period start measuring the metrics. It’s good practice and will give you insight into many things relating to the conversion rate, the success of the call to action and bounce rate. Once you’ve tested both versions you should have a clear picture of which one is most successful according to your pre-set metrics.
The Beginning
Where to start is dependent on the goals set for the website. Aims could be anything from generating sales to getting people to sign-up to a newsletter. It can be anything where the user’s behaviour is central to your aim. You can test many things using A/B Testing and the most commonly tested elements include:
- Headline descriptions
- The call-to-action of a page
- A sign-up or contact form
- Pricing pages/showcases
- Image placement
This is where you can begin and from this point you can follow our tips which give a clearer suggestion of what you should do to ensure good practice in your A/B Testing
Good A/B Testing Practice for Website Designers
A/B Testing takes time if you want to see results, so you must learn to be patient. Depending on the size of your project you can run A/B testing over a few days or many months and it’s highly dependent on the variables you’re looking at. Increasing the conversion rate of a site for example takes some time to generate results.
A/B Testing tools usually have an in-built feature which remembers where returning users have previously been and redirect them to that point. This ensures testing is consistent.
It’s also important to keep at it. If you carried out A/B testing once and didn’t get particularly helpful results, you should still continue to test. Over time improvements will mean that you reach your goals faster.
Variations need to be tested simultaneously to get a true picture. If you test different versions one after another it simply won’t work when using A/B. You need to split the traffic between your variations at the same time to see comparatively how they perform.
A/B Testing Practices to Avoid
First and foremost for clear results you must not include old visitors in new tests if the new website designs are significantly different to your current design. The risk of shocking and alienating regular customers is highly likely if you do this.
It is also important to trust the results of the test. Don’t go with your gut. There may be an element of the design you’re not keen on but if the customers are then you need to go with it. This website design is about the end user, not your personal preferences. A/B Testing is designed to show what works for the user, not what you like best.
Finally, you should remember that you can test a lot more than just design and aesthetic page elements. You can carry out A/B Testing of content to see how the end user responds.
A/B Testing Tools
Specialised software has been designed to make A/B Testing even simpler, below we’re giving two examples which could help your project.
Unbounce
Unbounce is designed for testing landing pages. Quick testing can be carried out and the software is highly intuitive and easy to use.
Optimizely
Optimizely is probably the best A/B Testing tool around. Highly sophisticated and packed with quality features – this is the tool to use if you’re committed to A/B Testing and want to begin immediately.
A/B Testing is a highly powerful tool for gauging how users are experiencing a website. It gives insight into how they’re using the site and how different elements are working or not working for them. When the time isn’t available to test hundreds of separate iterations, A/B Testing can be valuable too.
Website design can always be improved upon and there is always something more end users want. The more time you test and question a website the better your insight will be.
