Bulk Operations Endpoints
Create, update, and delete keywords and negative keywords in bulk through these endpoints.
Overview
Bulk operations accept arrays of items in a single request. All bulk endpoints share the same items array structure: each item carries a client-supplied correlationId and a data object containing the operation payload. The response returns a result array with one entry per input item, keyed by correlationId.
Find Bulk Endpoints
Each entity that supports bulk operations exposes create, update, and delete endpoints at the following paths:
Entity | Bulk Create | Bulk Update | Bulk Delete |
|---|---|---|---|
Keywords |
|
|
|
Negative Keywords |
|
|
|
Create, Update, and Delete
To create, update, or delete items in bulk, send a POST request to the entity’s bulk endpoint with an items array. The endpoint you call determines the shape of each item’s data object:
Operation | Description |
|---|---|
Create | Send the array to |
Update | Send the array to |
Delete | Send the array to |
Handle Errors
Bulk endpoints use a partial-success model when the request sets allowPartialSuccess: true. The endpoint processes valid items, and the response includes per-item error details for any items that failed. When allowPartialSuccess is omitted or false (the default), any single item failure rejects the entire batch.
The response reports bulk-level validation failures (such as sending more items than the endpoint allows in a single request) in the outer BulkResponse.error.details field, not in the per-item result array. When the endpoint returns a bulk-level error, it processes no items regardless of the allowPartialSuccess setting. If you hit this rejection, split the request into smaller batches and retry each one.
Choose Between Bulk and Single-Item Endpoints
Prefer bulk endpoints when creating, updating, or deleting more than a handful of keywords or negative keywords at once, such as rolling out a new keyword list across many ad groups or pausing a large negative keyword list in one pass. A bulk request counts as a single call against your rate limit regardless of how many items it carries, so batching changes into bulk requests is the most effective way to stay under the limits described in Applying Rate Limits during large-scale operations.
Use the single-item endpoints for one-off changes, such as adjusting a single keyword’s bid, where there’s no batch to build. The single-item response shape is also simpler to handle in application code, since it returns a single Error object instead of a per-item result array keyed by correlationId.