Personalizer Performance is an assessment of how effective the Personalizer is on your users. We essentially compare users who watch at least one recommended content versus those who do not. We try to see differences in these two groups to provide useful insights on the overall performance of the personalizer.
The consumption metrics below are the ones of interest. If these are better than the consumption of non-recommended content, the optimization is successful.
Percentage of the users who watched the recommended content offered by the JUMP Personalizer module endpoints.
This is calculated by the following equation: (Personalization #CA / Total #CA) x 100 = Personalization Reach %
Content Consumption Uplift
Change in monthly playback time amount (#hours) of all users who watched at least a recommended content over the users who did not watched recommended contents. This is calculated as the difference between the median monthly viewing time per user of those who have consumed recommendations and those who have not consumed recommendations. This difference is then multiplied by the # of unique users who watched at least one recommendation within a month.
Personalization Cumulative Users
The total number of subscribers or viewers, counted as unique, who watched at least a recommended content throughout a month, within a given time range.
Monthly Viewing Time
Average hours of content watched by the user in a month. Comparing those who watched the recommended content versus those who did not.
Note: These KPIs are offered on monthly basis aggregation and also provides the depiction of the metrics for the current month. It is not required to wait until the end of the current month to be able to depict them.