adidas performance marketing tests
As a member of the adidas Performance Marketing team, I tested our creative, audiences, and messaging at any chance I got. We were constantly learning and iterating. We measured performance, established a dynamic baseline, and gradually improved our campaigns from a true value standpoint.
See below for some examples of tests that we have run at adidas, along with their resulting recommendations and best practices.
Branded Content vs. Toolkit Creative: adizero Boston10
In order to stay up-to-date with creator trends we run AB tests and opportunistic non-true AB tests to see what type of branded content can outperform our toolkit creative. For this test we boosted 2 of our affiliate creators' adizero Boston10 Instagram posts simultaneously with the images that the adidas creative department produced, launched from adidas social.
Hypothesis:
The creators' content and audiences will result in a directional lift in performance against the toolkit creative.
​
​
Results:
​
-
Overall, the toolkit creative outperformed the Instagram Influencer branded content creative across most traffic metrics, most notably was a CPM 36-47% less than both influencer ads. It's likely the toolkit creative was successful due to the clear branding and product showcase, while the ad from @jessiezapo is more model focused and @goochybaby's ad features no model and cuts off the shoe. We typically see our best performance from more on-model and action shots with the product clearly visible.
​
-
The post from @Jessiezapo had the highest CTR at 3.04% and a CPC only 10% ($0.54) higher than the toolkit creative. It's possible that because the hero image is primarily of her, audiences were more inclined to click through the carousel to see more of the product itself.
-
​
-
Branded content ads likely garnered higher CPMs because they only ran on Instagram and IG Stories, whereas our toolkit creative ran against FB placements as well which gave the platform more placements to optimize against.​
​
Recommendations: Continue testing, including:
​
-
Utilizing products and branded content for different BUs
-
Consistent ad type and platform runs for branded content and toolkit creative
-
Variety of influencers (macro v micro)
Conceptual vs. On-Model vs. Product Only Creative: Stan Smith Retail Store Locator Test
Since we have observed that Conceptual vs. On-Model vs. Product Only creative assets perform differently, depending on the audience and platform, we wanted to test which of these creatives would perform best on our Retail Store Locator Display campaign, which should lead to an increase in aROI for this campaign.
Hypothesis:
Of Conceptual, On-Model and Product Only images, one will outperform the others in our Stan Smith Retail Store Locator Display campaign.
Results:
​
While impression delivery was fairly even across all 3 creative versions, the Product Only creative outperformed across the board, driving a stronger CTR, CVR, and overall, more revenue.
​
​Recommendations:
​
-
We can apply these learnings and use Product Only creative in our future Retail Store Locator Display campaigns, particularly in Q4 when we heavy up on spend, utilize our retargeting audiences, and drive the majority of our annual revenue
​
-
We launched our next Retail Store Locator Display campaign, using Product Only images for the Ultraboost21 line
UGC Carousel vs. Toolkit Creative: Formotion
Since we have observed that consumers are more likely to interact with products that are relevant to them, we wanted to drive Paid Social traffic to a broader bucket Product Landing Page and push up Formotion products and injecting a UGC Carousel.
​
Hypothesis:
The UGC Carousel should lead to an increase in engagement, and the effect will be measured by an increase in both quality traffic and conversion.​
Results:
​
-
Featuring a UGC Carousel for Formotion on Women’s Clothing New Arrivals PLPs after users from Paid Social clicked to site from a Formotion campaign post drove a +3% increase in Quality Traffic (the Primary KPI / >1 page view per visit), and in turn observed a validated decrease in Bounce Rate.
​
-
The UGC Carousel was also effective in driving an increase in PDP Visits, as well as directional increases in RPV & AOV
​
-
Conversion remained flat, however, most likely because shoppers served the UGC Carousel are purchasing higher-priced tights, rather than bras.
​
Recommendations:
​
-
Although presenting UGC for shoppers entering from Paid Social campaigns effectively increased Quality Traffic to adidas.com, we did not see this translate to Conversion. The best course of action from here is to pursue an end-to-end experience so that shoppers add relevant product to cart and continue on to purchase.
​
-
Because Paid Social traffic to this landing page was predominantly New Visits (80%), we should leverage product and user-experience that resonates most with new visits to further the Paid Social browsing journey on adidas.com.
