Learn how Blender gives you total control of your listings while still taking full advantage of Promoted.ai optimization.
Promoted’s “Blender” gives you control without interrupting optimization or limiting experimentation.
Using Blender, you can define:
- Allocation rules
- Quality scores
More than just manual rules like pinning and boosting, Blender supports advanced features like:
- Dynamic promotion or ad load
- Probabilistic mixing and spacing
Configure Blender using the configuration builder tool for ease and efficiency or edit the configuration directly in the Delivery SDK per Request for maximum control.
Use Allocation rules to insert, mix, space or diversify listings. Rules can be deterministic or probabilistic.
- Pin self-owned items at the top
- Mix-in slots for a competitive promotion or ad auction, if available
- Explore new contents and downrank spam
- Diversify listings by category or owner
- Mix formats like carousels and video with minimum spacing rules
Modify the Quality Score to change the ranking function. Promoted’s default quality score maximizes incremental conversions. You can use other ranking functions like recency or an in-house score, and you can combine these scores with Promoted’s scores.
Configure different quality scores definitions for different media types, like video playtime or carousel horizontal swipes.
Use Trimmers to remove items that exceed a score threshold, for example, when a quality score is too low or a spam prediction is too high.
Blender integrates with the Content Management System and data passed on Delivery API. Any attribute that you send to Promoted you can use in defining blender rules.
To optimize Blender rules themselves, use Promoted’s long-term value optimizer. For example, Promoted can learn the optimal mix of diversity, promotion load, and quality score definition to maximize future user retention and revenue.