The machine learning technology is looking at overall pacing to make sure impression goals are met and eliminate large fluctuations. The pacing is also optimizing towards when users are online and engaging. Online Audio is mainly about branding (Clicks may happen where companion ads appear, but it’s still mainly effective for branding), so instead of forcing to serve ads evenly throughout the day, the system optimizes based on the internet traffic patterns of users for when they are online and engaged.
This is what the machine learning system is looking at each minute as it is optimizing:
- Device type
- Browser
- Geo
- Audience Behaviors/Characteristics
- How frequently the user has been served impressions from this advertiser
- How recently the user has been served impressions from this advertiser
← Frequently Asked Questions