A/B Testing on forms

Gone are the days when you had to rely upon your graphic designer to tell you what looked good and worked online. You can now ask the most important people, i.e. your users. and in fact, not even ask them because they probably wouldn’t tell you what you needed to know; there’s a better way. Google Analytics and Google Optimizer. Used properly it’s the purest form of user-centred design.

User centred design is all about design being informed by the user behaviour and what actually works on the screen. In practice this is undertaken through A/B or multi-variate testing. A/B Testing is a great way to monitor and increase visitor conversions by placing a variant on an element or layout within your site and monitoring the click through rates. It’s easy enough to guess what sort of thing you think visitors will click on, but sometimes you will be surprised it’s the complete opposite – the only way to find out it to test.

Google randomises the variants and records which has the greatest clickthrough rate which in turn can be recorded against conversions. If you are testing more than one component at once this is known as multi-variable testing – the options available to test are limitless.

The elements that result in the highest percentage of conversions are the elements that you should continue to use on your site, you are then able to continue testing and improving other parts of your site too.

What sample size to test‌ Tests can be carried out over any period of time against any number of visitors, but there isn’t a right or wrong answer to say just how much time you should spend testing or to how many visitors; it completely depends on what you are testing, how much site traffic you get and what you are trying to achieve. Just make sure that you have tested a big enough sample that you can trust the result – if three is no obvious winner then you shouldn’t need to change your site unless you wanted a redesign, in which case you will be satisfied to know you can switch to your new design without affecting the conversion.

Forms that work

Don’t assume your visitors have any knowledge on filling in forms. Hold their hand all the way making it as easy as possible to fill in without being too patronizing ( you would be surprised how far this can be pushed). Simple rules to follow include:

  • Keep it simple
  • Make the text size big enough for anyone to read – they shouldn’t have to struggle
  • Questions need to be well written and very simple – if explanations might be needed put an ‘i’ or ‘‌’ icon at the end of the question
  • Make Field boxes big and obvious arrows if dropdowns
  • When possible use click graphics instead of typing fields
  • If visitors fail to fill in a required field it must be made very obvious to them which answer was missed – e.g red highlight around field or similar
  • The call to action button at the end of the form (or leading onto the next stage) should be one the most obvious elements on the page as for many it will be the conversion point.

The main thing to note when A/B testing is that the results will be different for almost every site you visit, so one site’s findings should not be the results you accept for your site. Every site must complete its own testing to be able to benefit from this exciting and proven method to benefit from increased visitor conversions.