Is it possible to test conversion over 2 pages vs 1 page?
I want to test two user journeys that utilise two variations of a widget at different page levels.
User is on homepage and uses widget variation 1
- Measuring if they click submit on widget
User is on homepage and clicks CTA in hero banner
User is on landing page and uses widget variation 2
- Measuring if the click submit on widget
Is this possible to set up? I've been through the URL redirect articles but only seen information that suggest you need to have the same number of pages in the experiment.
You may be misinterpreting the meaning of "the same number of pages".
The number of pages a user may visit can certainly be different, but from a Test Setup perspective, whatever "Pages" you add to the experiment will be part of the experiment in each variation (though some of them may not be accessible in certain variations).
It sounds like you only need to have 1 page defined for your experiment's Targeting - the homepage
Variation A - homepage: code that places Widget A on the page (or maybe it is already there)
Variation B - homepage: code that places the hero with the link to the landing page on the page (or maybe it is already there)
*If* your landing page requires modifications as part of the experiment, then you will need to create a Page for the landing page and add it to the experiment's Targeting.
Variation A - landing page: no code - it is expected that users in the A group will never even see this page
Variation B - landing page: code that places the widget onto the page
Analytics and Testing Guru
Running a "Conversion Rate" experiment where the paths are different lengths, is rarely a good indication of what is best for your business. I've run versions of this sort of test (small form with a submit button on the homepage vs a link to a page that has a similar form) many times, it is extremely likely that "Journey 1" will result in more Widget Submits than "Journey 2" (unless you are doing something horrible with the UX in Journey 1).
They key is not to measure which widget results in more Form Submits, but to instead measure the quality of the Form Submits. For example, consider 100 submits on the homepage with 12 quality leads vs 40 submits on the landing page with 12 quality leads. The variation with only 40 leads costs less time to work through than the variation that had 100 leads. Unless you care absolutely nothing about the cost of processing the leads, you'll want to work out a way of measuring the quality of leads coming through which each Variation (perhaps the variation identifier can be added to the widget's form so it is available on the backend)
Analytics and Testing Guru