Testing – General
Dare! Be provocative! Try to come up with variations that are significantly different from your original content. Experience shows that when the content variations are very similar to your original content results may be inconclusive, and the further away they are from each other, the higher the lift you will get in your conversion rates.
Keep in mind that this is an ongoing process: Once you have a winning version, you can go ahead and try out other content ideas and test them against your winning version.
Remember: this is only a test. The final decision on what content should be displayed on your website is yours.
Go against your gut feeling. Often the content variation that you feel will work best fails and the variation or combination that seems like it has no chance will produce the best results. So don’t be afraid to try out even those variations that your colleagues feel have no chance, or are way out there.
Not sure and want to see the versions before you show them to the world? Preview your test! As you set up you’re a/B or MVT test you can take advantage of Amadesa’s simulation tool, and preview your content variations to check out how the different element combinations work on your pages.
- Work with your creative team to prepare your different content versions.
- Coordinate with your analytics package. The Amadesa platform is able to integrate with a number of different analytics tools.
Use Multivariate testing when you want to learn about the influence specific page elements have on your users. When you have several elements on your page or funnel which you are considering as candidates for optimization, each with several alternative variations, Multivariate testing is the right tool to choose. It will require more traffic as a large number of combinations is tested but you will be able to get a more detailed analysis. A/B test is best to use on a page or funnel with less traffic, you will be able to create different user experiences each completely different than the other, replacing one or more page elements, but without an exact analysis of the influence each element had.
