Let's assume we a workflow with 3 steps.
Page 1 is our landing page.
Page 2 is the form (signup).
Page 3 is the conversion page (thank you page).
To test variations on page 1, I am running a multivariate test against several sections. There is a multitude of call to action buttons.
At the moment our conversion goal, is set on the thank-you page, and It means that the user clicked on the call to action, filled the form, and submitted correctly.
I feel that this approach is a bit dispersive, and doesn't tell me about the quality of the landing page.
Is it better to set the conversion on the call to action and keep the path shorter?