Skip to end of metadata
Go to start of metadata

Log Level Data Feeds

LLD feeds are not available via the API and require a custom cloud export configuration.

Log Level Data (LLD) feeds contain event-level data that enables you to build custom reporting and conduct in-depth analysis. Typically, log-level data is used for reporting on dimensions not available in the AppNexus Console reports. It can also be used for any analysis which requires event-level granularity, such as custom conversion attribution or identifying pools of high-value users.

To subscribe to Log-Level Data and learn about changes

If you are interested in receiving log-level data feeds, talk to your AppNexus representative about the data your use case will need, technical setup and pricing. 

On This Page

Overview

Log-level data feeds include information about the following:

  • Your managed publishers' transacted impressions. This includes: 
    • Information about the impression sold (e.g., publisher ID, tag ID, and domain)
    • Information about the user to whom the ad was served (e.g., user ID, geographic region, and user timezone)
    • ID of the creative that served
  • Your managed advertisers' transacted impressions, and any resulting clicks and conversions. This includes:
    • Information about the impression purchased such as advertiser ID, campaign ID, and creative ID
    • Information about the user to whom the ad was served (e.g., user ID, geographic region, user timezone, and number of times the user has seen creatives from your advertiser)
    • Information about the inventory where the impression served (e.g., Publisher ID or domain (if exposed by the seller))
  • Segment pixel fires for all of your network's owned segments (including network- and advertiser-level segments)

For a complete list of fields available to buyers and sellers, see the List of log-level Feeds section below.

Timestamps: If your log-level data is provided in the avro format or one of the protobuf formats, the timestamp field values will be supplied in UNIX Epoch time.

Use Cases

Typical uses for log-level data include:  

  1. Enhanced reporting - report on additional dimensions such as geographic region
  2. Segment management -  build segments based on user actions and report on first- party data if your user IDs can be mapped to AppNexus user IDs
  3. Conversion attribution - see all actions leading to a conversion, experiment with custom conversion attribution

Data Retrieval

Log-level data may be obtained via batch files. 

Use this method if data accuracy is paramount. Since batch files are relatively large, they cannot be consumed in the same way ordinary reports are. Clients who receive log-level data will typically have their own engineering resources or a third-party vendor build an ETL process to load the data into their own database. Once in the database, the log-level data can be used by their own reporting systems or data analysts. 

See Batch Log-Level Data for more information on how to intake your data.

Batch 

FeaturesBatch
Latencyover 1 hour
Viewability Information
Failure recovery
Export to Amazon S3
Data quality guaranteed
Recommended use cases
Financial calculations
Real-time dashboards 
Custom bidders 
Campaign performance metrics

Final, long-term analytics

LLD data is final and accurate. As a result, Batch LLD can be used to make financial calculations. If AppNexus discovers an inconsistency, Batch LLD data will be reprocessed. 

The table below lists which feeds are available via LLD, the feed handle for each and a link to view the feed columns.

 List of Log Level Feeds

FeedFeed HandleColumns
Standard Feedstandard_feed View
Segment Feedsegment_feedView
Bid Landscape Feedbid_landscape_feedView
Query String Feedcustom_data_feedView
Conversion Pixel Feedpixel_feedView
Video Event Feedvideo_events_feedView

Auction Categories Feed

auction_category_feed
View
Auction Segment Feed

auction_segment_feed

View
Key Value Auction Feedauction_kv_labels_feed
View
For sample files, see available downloads under the Format and Schemas section in Batch Log Level Data.

 

Related Topics