Twitter has introduced A/B Testing on the Ads API

author-image
Paawan Sunam
Updated On
New Update
Twitter Ads API


The new update for Twitter Ads API is designed to enable advertisers to generate ROI through optimised performance supported by A/B Testing.

The update to the Twitter Ads API gives advertisers the ability to conduct A/B testing to test the performance of variables such as bid type, creative, targeting, and more.

For advertisers and developers who participated in the beta testing of this product, no action is required. This launch expands access to the feature, but does not change existing workflows.

A/B Testing

This API-exclusive launch introduces a new capability for advertisers and creates the opportunity to better understand the levers that drive campaign performance.

Twitter is introducing a new endpoint (ab_tests) to be used for creating and structuring these tests. Within this endpoint, advertisers will be required to define:

  • The test duration, represented by the start_time and end_time fields
  • The level at which the split will occur, represented by the entity_type field
  • The number of user groups (min. 2, max. 30), each represented as an object in the user_groups array

Additionally, each user group must include the following information:

  • The percentage of users that should be allocated to the given user group, represented by the size field
  • The campaign IDs that should make up the pool of users for the given user group, represented by the entity_ids array

Also Read: Twitter sends legal notice to Musk; calls termination of acquisition invalid & wrongful

The A/B testing functionality is coming to the Twitter Ads API ahead of Twitter Ads Manager and will be exclusively available via the Ads API until Q4 2022he platform recommends that developers add support for the new ab_tests endpoint and associated parameters in order to uniquely support customers.

A/B Testing Setup

In order to successfully create an A/B test, advertisers must understand the following setup details:

Test duration (start_time, end_time)

The start_time and end_time values must be in the future (relative to when the A/B Test is created) and overlap with the included campaign flight dates.

The test must last at least one day for non-app-based campaigns and last at least five days for app-based campaigns.

A/B Tests can only be updated while the status is SCHEDULED. There is one exception: it’s possible to update the A/B Test’s end_time while it’s LIVE.

Split level (entity_type)

The entity_type must be set to CAMPAIGN. Though, we plan to also support LINE_ITEM in the future.

User groups (user_groups)

Each user group is represented as an object in the user_groups array with a minimum of two user groups required and a maximum of 30 user groups is allowed.

The size for each user group is set using a string representation of a numeric value between 1.00 and 99.00. The size values across objects must add up to 100.00.

The campaign IDs should be specified in each user group’s entity_ids array.

Optional Parameters

Advertisers will have the option to set the name and description for both the overall A/B Test and for each of the user groups.