Running an Experiment to test text ads versus call-only ads
If Experiments are set at the campaign level, meaning, I can only configure control/experiment elements that are at lower levels (i.e. keywords), how do I test campaigns? Specifically, I want to compare performances of different campaign types.
One is a call-only campaign and the other one is a standard text-ad campaign. The goal is to see which type has a better mobile conversion rate.
Re: Running an Experiment to test text ads versus call-only ads
Hi @Kevin P you'd have to "manually" split the coverage of the Campaigns, and there's a number of ways you could do that:
#1 - Ad Scheduling; apply schedules to the two Campaigns so that only one is running at any one time. Downside of this is the limits of the scheduling and difficulty in ensuring you get "fair" coverage (i.e. one Campaign isn't only showing from 3am to 7am).
#2 - Schedule by Script. You could use Scripts to pause/reenable Campaigns as often as once an hour, which would solve the limitations of the scheduling tool. You could even count impressions to ensure both Campaigns ended the day on the same number of impressions (although this would require both Campaigns to get a similar level of traffic).
#3 - Split by geo location. You could have one Campaign cover one set of States/regions and the other another. You'd have to try and ensure the areas covered included the same demographics/population.
Personally, I'd use #2, but then I'm a scripting guy
One word of caution: Have you considered that the text Campaign might be assisting the call-only Campaign? Running either/or might end up lowering performance on one or the other, or both.