You keep everything the same in the campaign and only change the landing page. That way, any difference in results is coming from the page, not the ad or the targeting. Most marketers split traffic between two pages and run the test long enough to get real data, usually at least a few hundred clicks. Changing too many things at once makes it impossible to know what caused a difference. So what do you think is the minimum amount of traffic needed before a landing page test gives reliable results?