How To Set Up A Creative Testing Structure
This blog was first published as a conversation in the Foxwell Founders Membership, and was too good to not share with the world. A special thank you to Phil Kiel and Jackson Cunningham for their help in collaborating on this article and providing such immense value for our members and now all of you reading this.
If you’re not sure if the membership is right for you and your advertising growth and continuing education, that’s ok!
Questions on whether the Foxwell Founders Membership might be a good fit for you? Or maybe you prefer learning and growing in silence and are interested in our courses? Shoot us an email at andrew@foxwelldigital.com and let’s chat.
Creative testing depends on a few things:
How much “excess fat” you have in your blended ROAS (every new asset you test is potentially going to reduce your return)
Do you need to improve performance? (do you need the new winning ads to be hit targets or are you already there?)
As a starting point, you could set a testing budget per day or per week, and then work out how many assets you could test with that budget. This will tell you how long it will take before you've tested every asset. For budgeting, start with all spend in testing, and then move to approximately 80% of TOF budget in evergreen and 20% in testing.
CBO Testing Campaign
Dynamic Creative ad set per concept (creative theme, hook, message etc.) So if it's: “This pill stops headaches,” that would be one concept in one ad set, if it's: “This pill relieves stress,” that's another ad set, and so on.
I wouldn't launch each ad set at the same time. Instead, stagger them every couple of days so you're not having to optimize and audit the performance of them all at the same time. Instead, you can check in over the course of a week on each of them individually.
Single ads ad set in the same campaign: This is where you'll pull out the Dynamic Creative assets that get the most spend or show the most positive signs (use the Post ID from the dynamic creative ad to retain some data against the creative)
For audiences, start with broad, then if broad isn’t profitable or successful, move to lookalike audiences or interest audiences for testing
CBO Evergreen campaign
Audiences you can rely on if you have them, if not then broad, LLA and interest stack ad sets
Winning ads from previous activity if you have them, if not this will be the place you duplicate the winning single ads if they continue to perform well in the single ads ad set above.
All winning ads go into every ad set, and turn them off if you're finding new testing ads that are beating the evergreen campaign, OR if the evergreen campaign as a whole is not profitable and hitting targets
When you’re ready to optimize, only turn off ads if the duplicate is performing better. We've seen it countless times when you duplicate an ad, it doesn't continue the same performance. For example, if you're getting a 3x ROAS from the testing DC, and a 2.5x from the single ad, don't pause them until you're getting the same in the Evergreen campaign, and even then, it may be best to leave all ads live if they're exceeding KPIs.
Note that creative testing can still include retargeting, BUT ultimately, the success of the ad account is driven by what happens at top of funnel. We also know that TOF campaigns will do some retargeting on their own, especially if you don't have maximum exclusions set up.
Like this content? Check out our membership – we have content like this every. single. week.