Testing Testing 123

Once analytics are used to understand customer behaviour and why thing happen, the next step in the optimisation process is to test the possible changes. Using tools like Optimizly, Qubit and Google Optimize we can help you find out how best to change your site to create a better user experience which will provide better conversion rates.

For Example, analytics is telling you that lots of customers are coming through to your landing page via your email campaign. However, once they get to the desired landing page they are not completing the action you want them to. This could be because the instructions of what to do next are not clear. It could be because they are struggling with the functionality of the site. It could also be because the offer is not compelling enough. You wont know which to change, if any need to be changed at all. You certainly don’t want to pay for development time only to find out the change didn’t work. This is where A/B & MVT come in.

A/B testing is where you test whole alternatives. So in out example you could test the landing page with a 25% off or a landing page with a free trail of your service. Once you get the results you know if your landing page should offer the free trail or 25% off.

MVT testing is were you test multiple combinations of changes in one test. E.G

Test One

    • New Content Version 1
    • Change CTA button To Green

Test Two

    • New Content Version 1
    • Change CTA button To Blue

Test Three

    • New Content Version 2
    • Change CTA button To Green

Test Four

    • New Content Version 2
    • Change CTA button To Blue

By testing multiple variations at the same time it can stop the need of making multiple iterations to and A/B test. You can also start to see the impact of each element you have changed, and the combined effect of some elements.

A/B it is better used when you have a complete idea of all the changes that could be made, why they should be made and how to make them with the ability to keep iterating to improve upon end result. For example if you done some customer research and the results suggest all the changes customers wanted to see.

MVT testing is best used when you are unsure of how to processed with website improvements and need some guidance. However, this mean it will take longer to reach statistical significance as you will have to split up your traffic up in order to test all the variations.

Want to have the ability to optimize your site and know the exact impact it’s going to have . . . get in contact.