Custom AI - Lookalike Audiences
收藏Snowflake2023-11-09 更新2024-05-01 收录
下载链接:
https://app.snowflake.com/marketplace/listing/GZT1ZGAPIN
下载链接
链接失效反馈官方服务:
资源简介:
Custom AI - Lookalike Audiences Overview:
Dstillery’s patented Custom AI modeling provides owners of first party data with a lookalike audience solution that takes precision and efficiency to the next level. Unlike traditional lookalike solutions that rely on segment overlaps to identify similar audiences, Dstillery’s Custom AI modeling scores and ranks over 300 million unique devices based on their similarity to your first party data. The modeling process yields 5 different output files with cutoffs across the precision-scale curve, so Snowflake users can choose the level of precision or scale necessary for their digital campaign.
All Dstillery Custom AI models utilize daily scoring technology combined with over a dozen patents that identify and suppress inaccurate data points to ensure you are reaching prospects most relevant to your business. Drive maximum performance for your advertising campaigns across any activation platform with Dstillery’s just-for-your-brand Custom AI lookalike audiences.
IDs supported:
* Dstillery ID
* UID2
* Hashed Emails (HEM)
Activation Platforms supported:
* The Trade Desk
* Adobe Audience Manager
* Nexxen
* DV360
* Xandr
* LiveRamp
Expected Workflow:
* Snowflake customer shares 1PD from customer Snowflake account to Dstillery account
* Inform Dstillery of activation platform along with Seat ID where LAL audience should get pushed
* Dstillery builds the LAL model using 1PD and then scores all browsers against the model
* LAL audience is generated as 5 segments distributed to your DSP Seat ID
* The 5 segments are labeled A-E where A is the highest precision / lowest scale and E is highest scale / lowest precision
* The Audience segments can then be targeted in any combination as needed
Tables Included:
* 1 Percent Model File (Highest Precision)
* 2 Percent Model File
* 5 Percent Model File
* 10 Percent Model File
* 20 Percent Model File (Highest Scale)
Fields Included:
* Segment ID
* Segment Name
* User ID
提供机构:
Dstillery
创建时间:
2023-11-09



