Audience Targeting for Glance Performance Campaigns

Glance, a consumer-first content platform, has access to over 400 million users worldwide. Every third Android device has Glance. These users consume diverse content on Glance, including news, health, lifestyle, food, sports, entertainment, automobiles, etc. This extensive user base provides valuable first-party data for targeted campaigns.

For performance campaigns, the Glance team uses these rich first-party data from its platform and other InMobi platforms as the foundation for the Glance Machine Learning models. Additionally, Glance also leverages advertisers’ data, real-time events, and past campaign data (advertiser-specific) to enhance targeting efficiency. The models dynamically create audience segments to target the right set of users.

Note:

Glance ads strictly follow data policies to safeguard advertisers’ data. The advertisers' and campaigns’ data are only used to scale their future Glance campaigns and are not shared with or used for other advertisers’ campaigns.

The data required to improve the efficiency of the models comes from the following sources:

User acquisition campaigns

For UA campaigns, Glance uses first-party and advertisers’ data to dynamically build the targeting rules and audiences:

  • Glance first-party data signals:
    • Consented data from Glance users and other InMobi platforms provide high-quality user behavioral, content consumption, and ad interaction data.
    • Real-time events data captured during the campaign (installs, post-install events, and user engagement) dynamically improves the targeting.
    • Past Glance campaigns provide insights into how users interact with various content and ad categories to precisely target high-value users.

      Note:

      Your data is only used to scale your future campaigns and is not shared with or used for other advertiser’s campaigns.

  • Advertisers’ unattributed data: You can leverage robust Glance ML models more efficiently if the models are trained on your unattributed data.  The data helps make the campaign targeting more precise, efficient, and compliant with user privacy regulations. Read more on how sharing unattributed data with Glance helps advertisers

Remarketing campaigns

For REM campaigns, Glance uses first-party and advertisers’ data signals to dynamically build the targeting rules and audiences:

  • Glance first-party data signals:
    • Consented data from Glance users and other InMobi platforms provide high-quality user behavioral, content consumption, and ad interaction data.
    • Past Glance campaigns provide insights into how users interact with various content and ad categories to precisely target high-value users.

      Note:

      Your data is only used to scale your future campaigns and is not shared with or used for other advertiser’s campaigns.

  • Advertiser’s first-party data signals:
    • Unattributed data can be used to train Glance ML models to target your preferred audience.  The data helps make the campaign targeting more precise, efficient, and compliant with user privacy regulations. Read more on how sharing unattributed data with Glance helps advertisers
    • Advertiser audiences are the set of audiences acquired by advertisers. Glance can show personalized ads to encourage re-engagement and conversions. Contact the Glance team to learn more about audience sharing.

Importance of unattributed audience data  

Although Glance first-party data are the primary source for building targeting rules and audiences, you must share unattributed data with the Glance team for the following reasons:

Improves ML model training and accuracy 

Unattributed data provide abundant information about user behavior and preferences and are not limited to ad-driven interaction. By enabling unattributed data on your MMP dashboards, you allow Glance to access a broader spectrum of data points. These include organic installs and user actions within the app, which contribute to more comprehensive training samples for Glance ML models.   

With more robust data sets, models predict user behavior more accurately and optimize ad placements effectively. 

Reduces exploration time on the open internet 

In the absence of unattributed data, Glance may allocate some campaign period and budget to identify potential conversion opportunities across various app categories. This process is costly and inefficient due to the high variance in conversion probabilities and the sparsity of conversion events. 

Instead, the Glance team can utilize the historical unattributed data patterns to make informed bidding decisions and allocate budget efficiently, thus, improving RoAS. 

Overcomes data sparsity challenges 

In scenarios where pre-historic data is sparse, unattributed data provides additional contexts and user interaction signals, even if it’s not directly linked to a conversion. This broader data view allows Glance internal systems to identify and learn from subtle patterns, user drops, and other signals that precede conversions, enriching the training dataset and enabling more precise targeting. 

On This Page

Last Updated on: 18 Oct, 2024