Target CPA Analysis – AppCase Study - Campaign
For one leading sports focused app, marketing executives were curious as to if YouTube could be a viable solution for their user acquisition efforts.
Our efforts proved educational and effect in obtaining an understanding of performance and scalability across 82 different campaigns. Over 4M views generated 30K clicks and 4.5K installs of our client’s sports app. Based on LTV data pulled from the advertiser’s app analytics the audience was rich enough to warrant an increase in our CPI threshold which we painted better performance in the latter part of the campaign.
Our largest challenge during this campaign was to find the best target audience & bid strategy with a core focus on conversion and CPI results. Moreover, geographic specific videos combined with a calendar schedule delivery strategy we were able to garner an integral understanding of performance.
- Audience Targeting
Six different targeting types tested: Broad, Custom Affinity, Topics, Look-A-Like Model, Affinity, ROI, and Retargeting
Audiences Tested: Males, 18-65+, Sports Fans, Tech Fans, Arts & Entertainment, Gamers, & Retargeting
- Bid Optimization
Two Bid Types: Manual & Target CPA
Key Metric Variables: CPV, CPM, CPC, CPI, & CTC
- Team Specific Targeting
Teams & Geo Targets Tested: 31
Mile Radius: : 5 mi, 10 mi, 50 mi
The solution Variable Media developed was a 16-week plan to spend $50K, drive results while more importantly testing into key challenge areas, team specific targeting, audience segments, and bid adjusting. Below reflects a monthly account of our campaign:
The launch of the campaign consisted of gathering key data points against our broad and audience segment approach, learning quickly that our sports affinity generated a 3x higher click to conversion and 65% more improved CPI.
CPI improved 27% from June while producing 10x more views at a lower cost per view contributing a higher CTR. Topics and affinity targeting types combined with our sports specific segments were the source of the vast month over month improvement in CPI. Geotargeting and team specific targeting also proved to be inefficient as we narrowed the audience price per view increased to levels causing an inflation of CPI.
- August & September
A turning point of the campaign was the introduction of target CPA bidding which improved our CPI month over month by 27%. Contrary to understanding when comparing a hyper-targeted audience against broad efforts under a CPA model, larger audience pools performed 21% better than targeted while also contributing an improved CPV, CTR, VR, and CTC.
Our sports app client was impressed with our results however with season’s end and playoffs in full swing, fewer users are going to games, and the campaign did not continue
Our efforts proved valuable insight in the major challenge areas
- Audience Targeting – Better than broad when implemented alongside a manual bidding strategy
- Bid Optimization – CPA Targeting proved highly useful precisely from a broad perspective limiting any audience strategy
- Team & Geo-Specific Targeting – Not the most efficient way to reach the target audience specifically from a cost perspective. The conversion was similar to our broad efforts, unfortunately. Moreover, geo-radius optimization proved ineffective as well.
By the numbers, the effort:
- Last nine weeks generated an average CPI 48% better than the first seven weeks
- Generated 4.4K installs, the majority (58%) of which in the latter part of the campaign which made up 45% of spend
- In September, the last month of the campaign all performance metrics garnered better than campaign average results, most importantly the lowest monthly CPI 21% more improved than the average.
- 39% fewer views can be sourced to our Target CPA efforts when compared to Manual bidding. Target CPA drove more installs contributing better performance results across the majority of metrics.