Question regarding the Mobile Cookbook Main Course (4)

Hey,
I have a question regarding the Mobile Cookbook Main Course...
In Part 2, it is written that I'm supposed to set up 4 campaigns for each angle. So 2 Angles = 8 Campaigns.
It is mentioned as well that I should split test between Affiliate Network as well for each campaign. Okay, Wouldn't that add up to 16 Campaigns if I split test between 2 networks? And if I want to test each campaign for Direct Linking and Landing Page, it'd add up to 32 campaigns: What da faq?
Am I just retarded or is it true? Or are some split tests not a must to put them into their own campaigns?

Thanks!

When you're split testing things like offers and landing pages, I'd do that in the same campaigns. Most trackers will allow you to set a specific weight for each URL/lander to be rotated.

If you broke up each offer and landing page into its own separate campaign on the traffic source, it's unlikely that all of them will be in the same spot in the ad chain (what position your ad appears in the rotation versus other competitors) which may skew your data and yield the results inconclusive.

I would however break things up at the campaign level for other reasons. Such as Android having an average CPM of $1 while iOS has $0.75. In that case I might split up my campaigns by OSs and bid accordingly on each one.

Unless you have a set of banners that is specific to each landing page, then you'd probably want to break those up into their own campaigns also. Or you could just set up redirect rules based on tracking tokens, all in the same campaign. For example, if the click came from banner A, load lander A, split offer A and B. Something like that. No right or wrong way.. Just do what works for you and gives you the data you need to optimize. As long as you know what combos produce the best ROI.

Originally Posted by integrity

When you're split testing things like offers and landing pages, I'd do that in the same campaigns. Most trackers will allow you to set a specific weight for each URL/lander to be rotated.

If you broke up each offer and landing page into its own separate campaign on the traffic source, it's unlikely that all of them will be in the same spot in the ad chain (what position your ad appears in the rotation versus other competitors) which may skew your data and yield the results inconclusive.

I would however break things up at the campaign level for other reasons. Such as Android having an average CPM of $1 while iOS has $0.75. In that case I might split up my campaigns by OSs and bid accordingly on each one.

Unless you have a set of banners that is specific to each landing page, then you'd probably want to break those up into their own campaigns also. Or you could just set up redirect rules based on tracking tokens, all in the same campaign. For example, if the click came from banner A, load lander A, split offer A and B. Something like that. No right or wrong way.. Just do what works for you and gives you the data you need to optimize. As long as you know what combos produce the best ROI.

Yep. If you want to split-test something it goes against the fundamental tenets of split-testing itself to add extra variables like campaigns. Rather it should always be done dynamically at some point in the chain without modifying other parts.