Case Study: Morning Complete
Activated You’s flagship product “Morning Complete” was having a hard spring.
The premium Health & Wellness product’s sales were low. ROI was below target and the content that drove a great January had burnt out on Facebook and Instagram.
The CPA on even the best performing ads had slipped to $10-$15 above target, with some performing far below.
I’ll take you behind the scenes of how I approached the problem, how content
strategy + content production worked hand in hand, and what our result was.
Step 1: Ramp Up Dynamic Creative Testing
A/B Testing vs. Dynamic Creative
DC testing is one of the most powerful tools on Facebook, but despite the brand occasionally using it for headline and text variations, they weren’t using it to it’s full potential for content, and were still testing their video with a slower A/B process.
My first goal was to introduce Dynamic Creative that took full advantage of the content variations possible (up to 10), come up with a template for designing future tests, work with editors to develop a workflow, and work with media buyers to teach the process we’d need to implement.
Step Two: Interviews & Soundbites
Producing and conducting on-camera interviews with actress Maggie Q, the face of the brand, my goal was to create a comfortable set and interviews that allowed her to speak comfortably, honestly about her health challenges - and create an authentic conversation.
Then, in the editing bay, I worked through the best moments of the interview to find authentic soundbites to build new content around.
Choosing the ones I believed were most compelling, I wrote scripts weaving together soundbites, on screen text, b roll, stock footage and original footage to tell a visual story.
Stress Effects
Nasty Comments
Stinky Poop
Great Day
The initial winner was “Stinky Poop,” in early testing with a CPA that was $10-20 below the target CPA, and $20-30 below the top performing ad at the time.
I had found multiple angles that looked promising, but my goal was to deliver enough content to let the consumer decide what made them want to spend their hard earned money - and then embrace the data which showed me what they wanted to inform the next set of content.
Step 3: 3-second strategy tweaks
Text variations
Imagery variations
Unique Combinations
Overseeing the whole process, and help guiding the hand of multiple creators means:
New text choices from copywriters
New visual choices from editors, graphic designers and animators
New combinations both planned and unplanned
All working together to provide the social channels with the glut of content required to take full advantage of machine learning.
Step Four: Data Diving
Spotting Patterns
Looking at the data begins from the day it’s uploaded and the ad is created, and continues for as long as it’s active -
What I found was the CPA’s were lower on images + text combinations that confused our audience, and less polished (read: uglier) opening images.
I was surprised - but I had data to prove it was working. It cut against creative assumptions, but I wasn’t afraid to follow the data to more content.
3 Second Views & ThruPlay
Identifying what were the best three second options, I began to look for new ways to iterate my own work and my collaborators in two directions
More “unusual combinations” of text and images,
Ugly foods and strange/weird food openings to see if we could enhance the trend.
Step Five: Fill the Pipeline
The info we’d seen from the slew of efficient, front sided content tests pushed my next rounds of creative in two directions -
Optimizing the existing Stinky Poop content - with tests on text placement, font colors, CTA variations and more to maximize conversions from those videos,
Ugly images of food performed better. So I started new scripts from scratch about a few “unusual” ingredients which would begin the process all over again from the top.
This messy picture of bitter melon turned out to be a financial winner.
