Intra-campaign experiments

Intra-campaign experiments are used to test a specific feature within a single campaign. Unlike system-managed experiments where traffic is split between control and treatment campaigns, intra-campaign experiments split traffic within the campaign, based on whether the feature is enabled or not.

This workflow is supported for the following ExperimentType values:

ADOPT_AI_MAX
Tests the impact of AI Max features on Search campaigns.
ADOPT_BROAD_MATCH_KEYWORDS
Tests the impact of broad match keywords on Search campaigns.
PMAX_TEXT_CUSTOMIZATION_FINAL_URL_EXPANSION
Specifically used for testing Performance Max features: text customization and final URL expansion. Enables Google AI to send traffic to relevant landing pages and generate text assets to better match search queries.

Setup

  1. Define the Experiment (with status set to ENABLED), providing an experiment type, a control ExperimentArm, and a treatment ExperimentArm:
    • For ADOPT_AI_MAX and ADOPT_BROAD_MATCH_KEYWORDS, both the control arm and the treatment arm must reference the same campaign in campaigns.
    • For PMAX_TEXT_CUSTOMIZATION_FINAL_URL_EXPANSION, only the treatment arm specifies the Performance Max campaign in campaigns (leave campaigns unset on the control arm).
  2. For ADOPT_AI_MAX, create a CampaignOperation update with a field mask to enable ai_max_setting.enable_ai_max and configure asset_automation_settings.

  3. Send a GoogleAdsService.Mutate request that includes mutate operations to create the experiment and experiment arms, and (for ADOPT_AI_MAX) to enable the test feature on the campaign.

For ADOPT_AI_MAX and ADOPT_BROAD_MATCH_KEYWORDS, traffic must be split evenly (50/50) within the campaign so that 50% of traffic is exposed to the enabled feature (the treatment group) and 50% is not (the control group). PMAX_TEXT_CUSTOMIZATION_FINAL_URL_EXPANSION supports custom traffic splits that sum to 100.

Java

// Create the experiment resource name using a temporary ID.
String experimentResourceName = ResourceNames.experiment(customerId, -1L);

// Create the experiment.
Experiment experiment =
    Experiment.newBuilder()
        .setResourceName(experimentResourceName)
        .setName("ADOPT_AI_MAX Experiment #" + UUID.randomUUID())
        .setType(ExperimentType.ADOPT_AI_MAX)
        .build();
MutateOperation experimentOperation =
    MutateOperation.newBuilder()
        .setExperimentOperation(ExperimentOperation.newBuilder().setCreate(experiment).build())
        .build();

// Create the control arm. Both arms in an intra-campaign experiment reference the same base
// campaign.
ExperimentArm controlArm =
    ExperimentArm.newBuilder()
        .setExperiment(experimentResourceName)
        .setName("Control Arm")
        .setControl(true)
        .setTrafficSplit(50)
        .addCampaigns(ResourceNames.campaign(customerId, campaignId))
        .build();
MutateOperation controlArmOperation =
    MutateOperation.newBuilder()
        .setExperimentArmOperation(
            ExperimentArmOperation.newBuilder().setCreate(controlArm).build())
        .build();

// Create the treatment arm.
ExperimentArm treatmentArm =
    ExperimentArm.newBuilder()
        .setExperiment(experimentResourceName)
        .setName("Treatment Arm")
        .setControl(false)
        .setTrafficSplit(50)
        .addCampaigns(ResourceNames.campaign(customerId, campaignId))
        .build();
MutateOperation treatmentArmOperation =
    MutateOperation.newBuilder()
        .setExperimentArmOperation(
            ExperimentArmOperation.newBuilder().setCreate(treatmentArm).build())
        .build();

// Create a campaign operation with an update mask to enable AI Max and configure asset
// automation settings.
// Note: For intra-campaign experiments, these settings are applied to the base campaign but are
// only active for the treatment traffic split.
Campaign campaign =
    Campaign.newBuilder()
        .setResourceName(ResourceNames.campaign(customerId, campaignId))
        .setAiMaxSetting(AiMaxSetting.newBuilder().setEnableAiMax(true).build())
        .addAssetAutomationSettings(
            AssetAutomationSetting.newBuilder()
                .setAssetAutomationType(AssetAutomationType.TEXT_ASSET_AUTOMATION)
                .setAssetAutomationStatus(AssetAutomationStatus.OPTED_IN)
                .build())
        .addAssetAutomationSettings(
            AssetAutomationSetting.newBuilder()
                .setAssetAutomationType(
                    AssetAutomationType.FINAL_URL_EXPANSION_TEXT_ASSET_AUTOMATION)
                .setAssetAutomationStatus(AssetAutomationStatus.OPTED_IN)
                .build())
        .build();

CampaignOperation campaignOp =
    CampaignOperation.newBuilder()
        .setUpdate(campaign)
        .setUpdateMask(FieldMasks.allSetFieldsOf(campaign))
        .build();
MutateOperation campaignMutateOperation =
    MutateOperation.newBuilder().setCampaignOperation(campaignOp).build();

// Send all mutate operations in a single Mutate request.
List<MutateOperation> mutateOperations =
    ImmutableList.of(
        experimentOperation,
        controlArmOperation,
        treatmentArmOperation,
        campaignMutateOperation);

try (GoogleAdsServiceClient googleAdsServiceClient =
    googleAdsClient.getLatestVersion().createGoogleAdsServiceClient()) {

  MutateGoogleAdsRequest request =
      MutateGoogleAdsRequest.newBuilder()
          .setCustomerId(Long.toString(customerId))
          .addAllMutateOperations(mutateOperations)
          .build();

  MutateGoogleAdsResponse response = googleAdsServiceClient.mutate(request);
      

C#

// Create the experiment resource name using a temporary ID.
string experimentResourceName = ResourceNames.Experiment(customerId, -1);

// Create the experiment.
MutateOperation experimentOperation = new MutateOperation()
{
    ExperimentOperation = new ExperimentOperation()
    {
        Create = new Experiment()
        {
            ResourceName = experimentResourceName,
            Name = $"ADOPT_AI_MAX Experiment #{ExampleUtilities.GetRandomString()}",
            Type = ExperimentType.AdoptAiMax
        }
    }
};

// Create the control arm. Both arms in an intra-campaign experiment
// reference the same base campaign.
MutateOperation controlArmOperation = new MutateOperation()
{
    ExperimentArmOperation = new ExperimentArmOperation()
    {
        Create = new ExperimentArm()
        {
            Experiment = experimentResourceName,
            Name = "Control Arm",
            Control = true,
            TrafficSplit = 50,
            Campaigns = { ResourceNames.Campaign(customerId, campaignId) }
        }
    }
};

// Create the treatment arm.
MutateOperation treatmentArmOperation = new MutateOperation()
{
    ExperimentArmOperation = new ExperimentArmOperation()
    {
        Create = new ExperimentArm()
        {
            Experiment = experimentResourceName,
            Name = "Treatment Arm",
            Control = false,
            TrafficSplit = 50,
            Campaigns = { ResourceNames.Campaign(customerId, campaignId) }
        }
    }
};

// Create a campaign operation with an update mask to enable AI Max and
// configure asset automation settings.
// Note: For intra-campaign experiments, these settings are applied to the
// base campaign but are only active for the treatment traffic split.
Campaign campaign = new Campaign()
{
    ResourceName = ResourceNames.Campaign(customerId, campaignId),
    AiMaxSetting = new Campaign.Types.AiMaxSetting { EnableAiMax = true }
};

campaign.AssetAutomationSettings.Add(new Campaign.Types.AssetAutomationSetting
{
    AssetAutomationType = AssetAutomationType.TextAssetAutomation,
    AssetAutomationStatus = AssetAutomationStatus.OptedIn
});

campaign.AssetAutomationSettings.Add(new Campaign.Types.AssetAutomationSetting
{
    AssetAutomationType = AssetAutomationType.FinalUrlExpansionTextAssetAutomation,
    AssetAutomationStatus = AssetAutomationStatus.OptedIn
});

MutateOperation campaignOperation = new MutateOperation()
{
    CampaignOperation = new CampaignOperation()
    {
        Update = campaign,
        UpdateMask = FieldMasks.AllSetFieldsOf(campaign)
    }
};

// Send all mutate operations in a single Mutate request.
List<MutateOperation> mutateOperations = new List<MutateOperation>
{
    experimentOperation,
    controlArmOperation,
    treatmentArmOperation,
    campaignOperation
};

MutateGoogleAdsResponse response = googleAdsService.Mutate(
    customerId.ToString(), mutateOperations);
      

PHP

This example is not yet available in PHP; you can take a look at the other languages.
    

Python

# Create the experiment resource name using a temporary ID.
experiment_resource_name = googleads_service.experiment_path(
    customer_id, "-1"
)

# Create the experiment.
experiment_operation = client.get_type("MutateOperation")
experiment = experiment_operation.experiment_operation.create
experiment.resource_name = experiment_resource_name
experiment.name = f"ADOPT_AI_MAX Experiment #{uuid4()}"
experiment.type_ = client.enums.ExperimentTypeEnum.ADOPT_AI_MAX

# Create the control arm. Both arms in an intra-campaign experiment
# reference the same base campaign.
control_arm_operation = client.get_type("MutateOperation")
control_arm = control_arm_operation.experiment_arm_operation.create
control_arm.experiment = experiment_resource_name
control_arm.name = "Control Arm"
control_arm.control = True
control_arm.traffic_split = 50
control_arm.campaigns.append(
    googleads_service.campaign_path(customer_id, campaign_id)
)

# Create the treatment arm.
treatment_arm_operation = client.get_type("MutateOperation")
treatment_arm = treatment_arm_operation.experiment_arm_operation.create
treatment_arm.experiment = experiment_resource_name
treatment_arm.name = "Treatment Arm"
treatment_arm.control = False
treatment_arm.traffic_split = 50
treatment_arm.campaigns.append(
    googleads_service.campaign_path(customer_id, campaign_id)
)

# Create a campaign operation with an update mask to enable AI Max and
# configure asset automation settings.
# Note: For intra-campaign experiments, these settings are applied to the
# base campaign but are only active for the treatment traffic split.
campaign_operation = client.get_type("MutateOperation")
campaign = campaign_operation.campaign_operation.update
campaign.resource_name = googleads_service.campaign_path(
    customer_id, campaign_id
)
campaign.ai_max_setting.enable_ai_max = True

for asset_automation_type_enum in [
    client.enums.AssetAutomationTypeEnum.TEXT_ASSET_AUTOMATION,
    client.enums.AssetAutomationTypeEnum.FINAL_URL_EXPANSION_TEXT_ASSET_AUTOMATION,
]:
    asset_automation_setting = client.get_type(
        "Campaign"
    ).AssetAutomationSetting()
    asset_automation_setting.asset_automation_type = (
        asset_automation_type_enum
    )
    asset_automation_setting.asset_automation_status = (
        client.enums.AssetAutomationStatusEnum.OPTED_IN
    )
    campaign.asset_automation_settings.append(asset_automation_setting)

client.copy_from(
    campaign_operation.campaign_operation.update_mask,
    protobuf_helpers.field_mask(None, campaign._pb),
)

# Send all mutate operations in a single Mutate request.
mutate_operations = [
    experiment_operation,
    control_arm_operation,
    treatment_arm_operation,
    campaign_operation,
]

response = googleads_service.mutate(
    customer_id=customer_id,
    mutate_operations=mutate_operations,
)
      

Ruby

This example is not yet available in Ruby; you can take a look at the other languages.
    

Perl

This example is not yet available in Perl; you can take a look at the other languages.
    

curl

Report on the experiment

Because control and treatment traffic are mixed within a single campaign, you must use direct experiment reporting to compare metrics between the control and treatment groups. Standard campaign-level reporting only shows aggregated metrics for the entire campaign and cannot distinguish between the two groups.

The following GAQL query retrieves click statistics for an ADOPT_AI_MAX intra-campaign experiment over the last 30 days:

SELECT
  experiment.resource_name,
  experiment.name,
  metrics.clicks,
  metrics.control_clicks,
  metrics.clicks_point_estimate,
  metrics.clicks_p_value
FROM experiment
WHERE segments.date DURING LAST_30_DAYS
  AND experiment.type = 'ADOPT_AI_MAX'

Promote, graduate, or end the experiment

After evaluating the results, you can complete the experiment using ExperimentService:

  • End: If you are not satisfied with the results, use EndExperiment. The feature is disabled, and the campaign reverts to serving all traffic without the experimental feature. This is a synchronous operation.
  • Promote (ADOPT_AI_MAX and ADOPT_BROAD_MATCH_KEYWORDS): If you are satisfied with the results of a Search intra-campaign experiment, use PromoteExperiment. This applies the experimental change as the new permanent state of the campaign. This is an asynchronous operation; see Asynchronous errors for details.
  • Graduate (PMAX_TEXT_CUSTOMIZATION_FINAL_URL_EXPANSION): PromoteExperiment is not supported for PMAX_TEXT_CUSTOMIZATION_FINAL_URL_EXPANSION; instead, use GraduateExperiment to apply the treatment arm's text customization and final URL expansion settings in-place to the existing Performance Max campaign. Include a CampaignBudgetMapping entry in campaign_budget_mappings specifying the treatment arm's campaign (experiment_campaign) and its budget (campaign_budget). This is a synchronous operation. (GraduateExperiment is not supported for ADOPT_AI_MAX or ADOPT_BROAD_MATCH_KEYWORDS because there is no separate treatment campaign to graduate.)