Implementing activation in your clean room¶
Overview of activation¶
The provider or consumer can send template results outside of the clean room in a process called activation. Snowflake supports three types of activation:
Provider activation, where results are pushed to a table in the provider’s Snowflake account.
Consumer activation, where results are pushed to a table in the consumer’s Snowflake account.
Third-party activation, where the provider or consumer pushes results to a Snowflake-approved third-party, such as LiveRamp or Meta Ads Manager, through an activation connector.
In all cases, the template must support activation the other party must approve activation for any columns of their own data that will be exported.
Activation supports differential privacy, if enabled, and respects differential privacy rules and budgets.
To activate data from a collaborator in a different cloud region, you must enable Cross-Cloud Auto-Fulfillment.
Provider and consumer activation¶
You can configure a clean room to save template results in the provider’s or consumer’s Snowflake account. Both the provider and consumer must approve activation of any data out of the clean room.
Activation is implemented using a dedicated activation template. In the clean rooms UI, an analysis template can be associated with an analysis template, and the user can run the analysis template and view the results, then run and activate the results of the associated activation template. The Snowflake-provided Audience Overlap & Segmentation flow does this.
An activation template need not be identical to its associated analysis template. The activation template is often a subset of the analysis template.
Supported templates¶
The following templates support provider and consumer activation:
Audience Overlap & Segmentation
SQL Query (UI only)
Supported combinations¶
Activation can be run either by the provider or the consumer. You can combine various combinations of provider- and consumer-run analyses with provider and consumer activation. (Learn more about provider-run analyses.)
The following combinations are supported:
Provider activation |
Consumer activation |
|
|---|---|---|
Provider-run |
✅ |
❌ |
Consumer-run |
✅ |
✅ |
Results¶
Provider activation results are saved to the provider’s account SAMOOHA_BY_SNOWFLAKE_LOCAL_DB.PUBLIC.CONSUMER_DIRECT_ACTIVATION_SUMMARY table.
Consumer activation results are saved to the consumer’s account in SAMOOHA_BY_SNOWFLAKE_LOCAL_DB.PUBLIC.PROVIDER_ACTIVATION_SUMMARY table.
See viewing results to learn how to read the data.
Implementing provider or consumer activation¶
Setup
Activation when using the clean rooms UI requires that the clean rooms account allows activation.
For provider-run activation, the clean room must be configured to support provider-run analysis.
1. Create or join a clean room
When creating or joining a clean room, in the Configure Analysis & Query step, under Activation Settings, specify which columns should be added to the results activated to your account.
2. Run the template and activate results
To run the activation associated with your analysis, complete these steps:
Run your analysis.
After running an analysis, select Results » Activate.
Under Activation Hub select the name of the provider or consumer account to activate to.
Provide information specific to the activation template, such as providing descriptive segmentation names or selecting activation columns.
Provide a segment name: this is an arbitrary string used to identify a set of results. You can provide a different string for each activation to group each run’s results separately, or you can use the same segment name over multiple runs with input variations if you want to combine those results.
Select Push Data.
See the next section to learn how to view the activated results.
Activation is performed differently depending on who runs it, and whether it’s consumer or provider activation.
Important
The first time a consumer activates data to a provider account in a clean room, the provider must establish a data pipeline by signing in to the clean room UI for that account and staying signed in for up to 30 minutes. This needs to be done only once per clean room per consumer. Until that is done, data will not appear in the provider’s account, even if the activation succeeds.
Here is how a consumer can push results to their own Snowflake account.
Provider
Create the clean room, link datasets, and set join policies, as for a standard clean room.
Either choose a supported Snowflake standard template, or add a custom activation template to the clean room. If this clean room is to be used in the UI, you must provide a web form with the proper activation fields, as described in the template documentation.
Enable the template for consumer activation by calling
provider.enable_template_for_consumer_activation.Set the activation policy in the clean room for that template to specify which provider columns can be activated. All columns used in the activation template must be in the activation policy, whether they are projected or not.
Add consumer collaborators, set the default release directive, and publish the clean room, as usual.
Consumer
Install the clean room, link datasets, and set join policies, as for a standard clean room.
Set the activation policy in the clean room for that template to specify which consumer columns can be activated. All columns used in the activation template must be in the activation policy, whether they are projected or not.
Run the activation by calling
consumer.run_activation, with the last parameter set to TRUE to indicate a consumer activation.View the results, as described below.
Examples
Download the following examples and upload them as worksheet files in your Snowflake account. You will need separate accounts for the provider and consumer, each with the clean rooms API installed. Replace the information as noted in the sample files.
Here is how a consumer can push results to a provider’s Snowflake account.
Important
If the consumer and provider don’t both have the clean rooms UI installed, and the consumer is activating to the provider:
The consumer must run the following SQL command:
ALTER SHARE SAMOOHA_INTERNAL_GOVERNANCE_SUMMARY_SHARE_NAV2 ADD ACCOUNTS = $provider_account_data_sharing_id;
where
$provider_account_data_sharing_idis the provider’s Data Sharing Account IdentifierThe provider must run the following procedure:
CALL samooha_by_snowflake_local_db.provider.mount_provider_activations_share( $consumer_account_data_sharing_id, TRUE, FALSE);
where
$consumer_account_data_sharing_idis the consumer’s Data Sharing Account Identifier.
1. Provider
Create the clean room in the standard way.
Link datasets. The provider must also link the
samooha_by_snowflake_local_db.library.temp_public_keytable into the clean room.Set the join policy in the standard way.
Either choose a supported Snowflake standard template, or add a custom activation template to the clean room. If this clean room is to be used in the clean rooms UI, you must provide a web form with the proper activation fields.
Set the activation policy in the clean room for that template to specify which provider columns can be activated. All columns used in the activation template must be in the activation policy, whether they are projected or not.
Add consumer collaborators, set the default release directive, and publish the clean room, as usual.
2. Consumer
Install the clean room, link datasets, and set join policies, as for a standard clean room.
Set the activation policy in the clean room for that template to specify which consumer columns can be activated. All columns used in the activation template must be in the activation policy, whether they are projected or not.
Run the activation by calling
consumer.run_activation, with the last argument set to FALSE to indicate a provider activation.
3. Provider
The first time a consumer activates data to your account you must sign in to the clean rooms UI for this account for about 30 minutes after the consumer has activated data. After that, the data will appear in your account. This is done only once per clean room per consumer account. Later activations by the same consumer in the same clean room do not need this step.
View the results, as described below.
Examples
Download the following examples and upload them as worksheet files in your Snowflake account. You will need separate accounts for the provider and consumer, each with the clean rooms API installed. Replace the information as noted in the sample files.
Here is how a provider can push results their own Snowflake account. This combines several techniques, including custom templates, provider-run analysis, and provider activation, and so involves several rounds of request and approval between the provider and consumer.
1. Provider
Create the clean room, link datasets, and set join policies, as for a standard clean room, with one exception: You must link in the table
samooha_by_snowflake_local_db.library.temp_public_key. Provider-run data is encrypted, and this enables encryption and decryption of the results.Either choose a supported Snowflake standard template, or add a custom activation template to the clean room. If this clean room is to be used in the UI, you must provide a web form with the proper fields to support activation, as described in the template documentation.
Set the activation policy in the clean room for that template to specify which provider columns can be activated. All columns used in the activation template must be in the activation policy, whether they are projected or not.
Add consumer collaborators in the standard way.
Enable provider-run analysis in the clean room by calling
provider.enable_provider_run_analysis. This must be done after adding collaborators but before collaborators install the clean room. If you change this setting after a consumer installs the clean room, the consumer must reinstall the clean room for the change to take effect.Set the default release directive and publish the clean room, as usual.
2. Consumer
Install the clean room, link datasets, and set join policies as in a standard clean room.
Set the activation policy in the clean room for that template to specify which consumer columns can be activated. All columns used in the activation template must be in the activation policy, whether they are projected or not.
3. Provider
Request permission from the consumer to run your activation template by calling
provider.request_provider_activation_consent.
4. Consumer
Grant the provider permission to run a given template in this clean room by calling
consumer.enable_templates_for_provider_run.Grant the provider permission to activate results from a given template in this clean room by calling
consumer.approve_provider_activation_consent.
5. Provider
Enable consumer data to be shared in a provider activation by calling
provider.mount_request_logs_for_all_consumers.Run the activation template by calling
provider.submit_analysis_request). The request takes several minutes to appear in the logs; check status by callingprovider.check_analysis_status. Note that even after status is reported as SUCCESS, additional time is required for results to be decrypted and written to the provider’s Snowflake table. All decrypted data is appended at one time to the results table. Keep checking the results table periodically for your segment or activation ID.Note
To modify a template after the consumer approves it, you must take the following steps, or else
provider.submit_analysis_requestwill continue to run the last approved version of the template.
Provider updates the template by calling
provider.add_custom_sql_template. No need to callcreate_or_update_cleanroom_listingagain.Consumer calls
consumer.enable_templates_for_provider_run.Consumer calls
consumer.approve_provider_activation_consent.The updated template is now ready for provider activation.
Common errors
Object cleanroom_name.CLEANROOM.TEMP_RESULT_DATA does not exist or not authorized- Temporary results table could not be generated for some reason. Could be a SQL error in the template, or your template didn’t explicitly generate a table; look at the error details.Query validation checks failed- Some columns used in the template that weren’t in the activation policies.
Examples
Download the following examples and upload them as worksheet files in your Snowflake account. You will need separate accounts for the provider and consumer, each with the clean rooms API installed. Replace the information as noted in the sample files.
Viewing provider and consumer activation results¶
Activation results location and format¶
All provider activation results are appended to a Snowflake-designated table. Each row in the table maps to a row in the query result. Additional query results are appended to the table. You can distinguish between different runs by the ACTIVATION_ID column, which is unique per activation, and the SEGMENT column, which is specified by the caller for each activation run.
Provider activation results are stored in SAMOOHA_BY_SNOWFLAKE_LOCAL_DB.PUBLIC.PROVIDER_ACTIVATION_SUMMARY in the provider’s account.
Consumer activation results are stored in SAMOOHA_BY_SNOWFLAKE_LOCAL_DB.PUBLIC.CONSUMER_DIRECT_ACTIVATION_SUMMARY in the consumer’s account.
These tables contain the following columns:
- USER_ID:
One row of results, in JSON format, where the keys are the column names and the values are the value for that column in that row. The object also contains a column for each argument passed into the template.
- ACTIVATION_ID:
A unique ID for each request. The ID is returned from a successful activation request. You can filter by this column to get all results for the same activation run, or filter by SEGMENT if you reuse the same segment name across multiple runs. This is the same as the query request ID returned by
submit_analysis_requestorrun_activation.- CLEANROOM_NAME:
Name of the clean room where the query was run.
- CONSUMER:
(Provider activation only) The consumer who approved this activation.
- PROVIDER:
(Consumer activation only) The provider who approved this activation.
- SEGMENT:
An arbitrary string value that you assign when you run the activation. This column enables you to join results across multiple query runs.
- TIMESTAMP:
When the activation was run.
Provider activation example
SELECT * FROM SAMOOHA_BY_SNOWFLAKE_LOCAL_DB.PUBLIC.PROVIDER_ACTIVATION_SUMMARY WHERE segment = 'my_segment';
USER_ID | CLEANROOM_NAME | SEGMENT | CONSUMER | TIMESTAMP | ACTIVATION_ID
"{""AGE_BAND"":55,""ITEM_COUNT"":2328,""STATUS"":""MEMBER""}" | test activation | my_segment | ABC1234 | 2025-04-01 16:27:14.068 | cleanroomactivationdataanalysisresults20250401231728469
"{""AGE_BAND"":20,""ITEM_COUNT"":88,""STATUS"":""PLATINUM""}" | test activation | my_segment | ABC1234 | 2025-04-01 16:27:14.068 | cleanroomactivationdataanalysisresults20250401231728469
"{""AGE_BAND"":80,""ITEM_COUNT"":18,""STATUS"":""GOLD""}" | test activation | my_segment | ABC1234 | 2025-04-01 16:27:14.068 | cleanroomactivationdataanalysisresults20250401231728469
...
Reading provider or consumer activation results¶
Run the appropriate SQL command to view results activated to your Snowflake account:
View provider activation results
SELECT *
FROM SAMOOHA_BY_SNOWFLAKE_LOCAL_DB.PUBLIC.PROVIDER_ACTIVATION_SUMMARY
[WHERE segment = <SEGMENT_NAME>] [AND activation_id = <ACTIVATION_ID>];
View consumer activation results
SELECT *
FROM SAMOOHA_BY_SNOWFLAKE_LOCAL_DB.PUBLIC.CONSUMER_DIRECT_ACTIVATION_SUMMARY
[WHERE segment = <SEGMENT_NAME>] [AND activation_id = <ACTIVATION_ID>];
Each row of data is combined into an object in the USER_ID column. You can flatten results using a query like the following:
-- Assuming columns AGE_BAND, STATUS, and ITEM_COUNT
SELECT
item:"AGE_BAND",
item:"STATUS",
item:"ITEM_COUNT"
FROM (SELECT parse_json(user_id)
AS item
FROM SAMOOHA_BY_SNOWFLAKE_LOCAL_DB.PUBLIC.PROVIDER_ACTIVATION_SUMMARY
WHERE segment = $segment_name)
ORDER BY item:"AGE_BAND", item:"STATUS" ASC
LIMIT 20 ;
View the latest 10 result rows in Snowsight:
Open Snowsight and navigate to Data » Database.
For provider activation navigate to
SAMOOHA_BY_SNOWFLAKE_LOCAL_DB»PUBLIC»Tables»PROVIDER_ACTIVATION_SUMMARY.For consumer activation navigate to
SAMOOHA_BY_SNOWFLAKE_LOCAL_DB»PUBLIC»Tables»CONSUMER_DIRECT_ACTIVATION_SUMMARY.Select Data Preview.
Third-party activation¶
Third-party activation deposits query results in the account of a Snowflake-approved third party using a third-party activation connector.
Third-party activation is supported only in the clean rooms UI, and not using custom templates.
Activation when using the clean rooms UI is supported only if the clean rooms account allows activation.
The clean rooms administrator must configure the environment to support third-party activation connectors, select the allowed connectors, and configure them, before they can be used in any clean room.
Third-party activation supports both consumer- and provider-run analyses.
Supported templates¶
The following templates support third-party activation:
Audience Overlap & Segmentation
SQL Query
Implementing third-party activation¶
Create or join the clean room: When creating or joining the clean room, in the Configure Analysis & Query step, under Activation Settings, specify which columns should be added to the results activated to your account.
Activate results:
Run your analysis.
After running an analysis, select Results » Activate.
Under Activation Hub select the name of the third-party provider to activate to.
Provide information specific to the provider. This can be providing descriptive names or selecting which columns to activate. The tooltips on the page should provide additional information for that provider.
Select Push Data.