How Facebook’s purchase event impacts instant donor conversion
Care Net
Experiment Summary
Timeframe: 01/24/2019 - 02/25/2019
Care Net offers people the opportunity to get a free devotional called, Prayers for Life. This offer is promoted through paid Facebook ads. In previous testing with Care Net’s online courses, we had found that the purchase event in Facebook significantly increased instant donor conversion. Because the Prayers for Life offer was a different kind of offer than the online courses where we saw that success we wanted to test this approach with this kind of offer as well to see if we’d get similar results. This campaign was originally set up using Facebook’s custom conversion which was based on an event we manually set up. We tested this approach against Facebook’s native purchase event.
Research Question
What kind of impact would Facebook’s purchase conversion have on instant donor conversion?
Design
Results
Treatment Name | Conv. Rate | Relative Difference | Confidence | |
---|---|---|---|---|
C: | Custom Conversion Optimization | 0.01% | ||
T1: | Purchase Optimization | 0.09% | 517.8% | 100.0% |
This experiment has a required sample size of 5,978 in order to be valid. Since the experiment had a total sample size of 56,645, and the level of confidence is above 95% the experiment results are valid.
Flux Metrics Affected
The Flux Metrics analyze the three primary metrics that affect revenue (traffic, conversion rate, and average gift). This experiment produced the following results:
0% increase in traffic
× 517.8% increase in conversion rate
× 0% increase in average gift
Key Learnings
By using Facebook’s purchase conversion, we were able to significantly impact instant donor conversion. In fact, this conversion event increased instant donor conversion by 517.8%! The results of this experiment dramatically changed the method in which we view and execute paid acquisition campaigns moving forward.
The purchase conversion allows Facebook to pull in additional purchase data to optimize the campaign instead of relying on a singular conversion metric that we tell it to do.
Question about experiment #13267
If you have any questions about this experiment or would like additional details not discussed above, please feel free to contact them directly.