Optimization testing in Kentico
December 29, 2016
By: Petar Kozjak
In: AB Test, Kentico EMS
I had lot of scenarios where a site is launched and the client has high expectations on generating new leads. PPC campaign is on the way, a fancy new landing page is created but results are not as expected. First problem that I saw was that my client thinks he knows his customers. Well does he really?
Knowing your customers and what they want is not that simple. Here are few arguments that I hear a lot:
- Our company color is this and we need to use it
- Our customers will love this content
- This video is perfect
- We like short forms, so do our customers
- Short CTA´s are good looking
My argument on all of this is: Do you want to generate new leads or impress visitors with colors, buttons, content and videos?
We can argue the whole day who is right or wrong and still not get it done. There is only one simple way to find out what your customers want and that is: Optimization testing. We will perform a set of tests and based on the results decide what gets us better results.
There are two type of tests: A/B Tests and MVT Tests. The difference between the two is simple. In A/B we test two different versions and in MVT we test only variants of elements like video, title, form.
Best way is to start with A/B Test and then later on with MVT. If you start with MVT you could end up with a lot of different variations. That will just create a large amount of data that needs to be validated.
Kentico allows us to easily setup A/B or MVT tests. In this blog, I will not show an example because the Kentico documentation is very good but I will give you some guidelines on what to look for in your tests.
Hints for optimization testing:
- Start with A/B test
- Define a time span for your test
- Create a different version of form. There is large argument whether to use short or long forms. Again, there is no universal answer.
- Use a real button (CTA). My advice would be to create a large one with a strong color that looks like a button. I do not like ghost buttons, especially on landing pages, but feel free to test one.
- Play with your headline, few word switches could make a huge difference
- Change layout
- Change colors
- Analyze your data
- If you are still not 100% sure for some elements start with MVT
Where to start:
A/B testing website pages
The Importance of A/B Testing
The Importance of A/B Split Testing to Improve the Performance of Your PPC Campaign
Comparing a Multivariate Test to an A/B Test