You can adjust all of your cookie settings by navigating the tabs in this window.
When it comes to marketing/advertising, it’s often the little adjustments that lead to enormous changes (whether positive or negative), like shifting tectonic plates. Looking at performance marketing on Facebook, Nanigans recently constructed a study within the ecommerce vertical highlighting Facebook “targeting types” as being able to have such an impact on campaigns.
Working with multiple ecommerce customers, the nature of the study was specifically around using targeting to scale campaigns while keeping the average cost per customer price under $20.
Seven different targeting types were tested throughout the study; results – based on data from Nanigans’ clients – for the first six targeting options were as follows:
None: As expected, without any specific targeting, cost-per-purchase rates were exorbitantly high.
Partner Categories: A step in the right direction comparatively against “no targeting” but costs were still high hindering scale.
Lookalike: With general lookalike targeting, volume started to increase while cost-per-purchase decreased, however, metrics were still too high to scale.
Broad: With broad targeting employed, volume increased and the cost-per-purchase was close to goal. Ultimately, however, this segment didn’t scale at performance goals
Custom Audiences: Results with custom audiences were encouraging as the cost per purchase was under goal, however, generating the needed additional volume proved difficult.
FBX: While FBX retargeting should have produced the most efficient cost-per-purchase, which it did, the volume was especially exciting to see.
With the results of the preceding ad options, it became evident that a mix of Broad, Custom Audience and FBX targeting would meet our overall campaign goals, however, we still weren’t at the level of scale that we wanted to be for the test.
…until Interest Targeting…
Interest Targeting: Campaign expectations and spend projections changed almost instantaneously after launching interest based targeting.
Excluding FBX, the highest impression level we were able to reach within goal was 197,033,385 leveraging Customer Audiences.
With Interest Targeting, however, impression delivery increased to 7,423,499,277, a 3,667% increase. The important metric to focus on, however, is the cost-per-purchase which averaged $11.89, an improvement of 8% compared to Custom Audiences. The purchase rate did decrease from 5.29% to 4.08% representing a 22% decrease in efficiency; however, this is not atypical when scaling campaigns.
Takeaway for Marketers: When running a performance marketing campaign on Facebook at scale, gather as much data as you possibly can. Targeting matters, and makes all the difference between running a campaign that can deliver billions of impressions (that drive ROI) instead of just millions.