Creating Experiments

Multivariate testing is not as complicated as it sounds. This article will help you perform the best experiments with Twik, here’s how.

Twik is an optimization system which allows AB and multivariate testing with a various array of elements. Before we get into the creation of new experiments, here is a list of Twik types

1. Anchor

Change your call-to-action or category name. Anchor Twiks allow you to change the text of a link – be it a category on the menu, a call-to-action, or a simple on-page URL.

2. Final URL

Change the traffic destination. URL Twiks allow you to test the best funnel for your users by changing the directed URLs wherever you choose.

3. Image

Customize visuals for different visitors. Image Twiks allow you to change the images on your site to fit the preferences of various users allowing you to find the best image for each type of user.

4. Form

Modify forms to reach best practice of maximum conversions. Form Twiks will help you change the value or placeholder of the form field.

5. Iframe

Change embedded videos or iframes.  Iframe Twiks allow you to easily change the content of your iframe.


Try out new markup on your page. Test the effects of new page layouts and properties on your conversion rates.

7. Script

Insert scripts to your page to make custom changes. Personalize your website with a more complex logic – The sky is the limit.

Now that we got that out of the way, here is a short experiment guide.

How to create a new experiment:

  1. Tag the element you wish to modify –  a pop-up will appear where you will need to name the Selector
  2. Define Audience – choose from various demographics and targeting options to define the best audience for your experiment.
  3. Define variants – variants can be in a paused or active state, containing the values of the tests. For example, if you wish to change your call-to-action from “Sign Up” to “Registration”, your variant will contain the phrase “Registration”.
  4. Preview
  5. [Optional] Form Groups

Twik uses A/B and multivariate testing in order to find the best performing combination of variants. If you need to control the combinations of variants, “Groups” is the way to go.

For example, let’s say you want to test the display of two different products on your homepage. For this purpose, you need to create two variants of each image, text, and URL. The system default mode is multivariate testing, meaning that sometimes the product image can appear with the wrong text or URL. To prevent this situation from happening, we use groups. In this case, we will create 2 groups, one for Product A and one for Product B. This will limit the system to test one group at a time.

Forced Group – Imagine a situation where you want to have a complex test on your homepage, like reorganizing components and changing titles. The thing is, you want to have the same order in every test and only change the title. In this case, you need to create a forced group, a situation where every test must use different variants from each selector.

Was this article helpful?