Migrate to the OpenRTB protocol

Google will migrate from Google Authorized Buyers protocol to the OpenRTB protocol on 2/15/2025 to align more closely with industry standards. This should ultimately reduce integration friction and maintenance complexity. We are providing this migration guide to assist you with transitioning to the OpenRTB protocol.

This guide consists of the following:

Mapping guide for handling things like fields, taxonomies, and extensions
Experimentation information
How to access technical support and/or address your questions

Mapping Guide

Visit the developer site for the full mapping guide. If you have feedback, visit the Technical Support section or share your feedback with your account manager.

Experimentation

For your accounts, Google can set up an A/B experiment to test the performance between an OpenRTB endpoint and a Google Authorized Buyers Protocol endpoint. Reporting will be available through a cloud bucket. Follow the steps below to set up an A/B experiment.

Set up an experiment

Set up your OpenRTB serving infrastructure and provide the URL to Google by completing this form.
Google creates the bidding endpoint configuration in the Authorized Buyers user interface.
On receipt of the form, Google team will activate the experiment within 3 days.
Once the experiment has been activated, the traffic is split at the specified percentage between Google RTB production endpoint and the test OpenRTB endpoint.
You receive key metrics and reporting via the publisher settings storage bucket. This should include any major fields or metrics for performance monitoring.

Note: Bidders can start experiment traffic as low as 0.1% and go as high as 20%. They can contact Google to change the percentage of traffic affected by the experiment and when they are ready to launch/dial up the experiment traffic.

Metrics to monitor	Adverse behavior	Potential root causes
Number of bids	Decrease	Problems with bid request processing on bidder’s end (for example, unable to ingest specific signals such as category/attribute taxonomies).
Bid values	Decrease
Number of response parsing or processing errors	Increase	Issues with bidder’s response building Examples for some common issues: The creative pertains to categories blocked by the publisher. Note that OpenRTB uses a different category taxonomy than Google RTB. See the new detectedCategories field in the CreativeServingDecision. Ineligible billing ID for deal. See the list of eligible billing IDs in `BidRequest.imp.pmp.deal.ext.billing_id`. Invalid event notification token, which must be a UTF-8 string, unlike in Google RTB where arbitrary byte strings are supported.
Number of bids filtered	Increase
Latency	Increase	HTTP reconnects given HTTP connection is more likely to expire on experimental endpoint. OpenRTB logic is taking longer than Google RTB logic. Degradation in performance possibly caused by an increase in the time required to process the request and build a response.
Number of timeouts	Increase

Note: If you want to stop or disable the experiment, set the QPS of your experimental endpoint to zero. Once you’re ready to progress past testing, you can move all queries to the OpenRTB endpoint.

Technical Support

For technical support or questions about the process, here are some options:

Contact your sales representative or raise a customer ticket to help with account configuration, product and UI questions, and some technical support. If you don’t have an assigned point of contact, use the contact form.
Access Contact Us to raise a customer ticket.
Authorized Buyers support forum provides public answers about the RTB protocols or APIs. You can search for similar questions to find answers to your questions.
Use the Technical API email support for answers to private questions about the RTB protocols or APIs. Unlike the support forum, questions here are private and can include details about your account that may assist with investigating your question or issue.

Was this helpful?

How can we improve it?