Lift reporting

To measure the causal impact of your ad campaigns on brand perception, conversions, conversion value, and organic search behavior, query the lift measurement reporting resources in the Google Ads API:

  • Brand Lift reporting: Evaluates how ad exposure influences brand perception across surveyed measurement types in Google Ads API v25 and later.
  • Conversion Lift reporting: Measures incremental conversions, conversion value, cost per incremental conversion (iCPA), and incremental return on ad spend (iROAS) against an unexposed holdback group in Google Ads API v25 and later, including:
  • Search Lift reporting: Quantifies the incremental increase in organic search query volume across Google Search and YouTube driven by ad exposure in Google Ads API v26 and later.

For standard A/B experiment reporting on the experiment and campaign resources, see Report on experiments.

Brand Lift reporting

To evaluate how your ad campaigns influence brand perception in Google Ads API v25 and later, query the dedicated Brand Lift measurement resources.

Brand Lift resources and breakdown views

Brand Lift configurations, flights, and breakdown views are read-only in the Google Ads API (all fields are OUTPUT_ONLY) and are configured in the Google Ads UI or with your Google representative or GPS contact. Query these resources by using the customer_id parameter of the account (manager or child account) that owns the study.

Query the LiftMeasurementFlight resource to retrieve flight-level Brand Lift schedule and survey collection metadata, including the flight status (lift_measurement_flight.status, defined by the LiftMeasurementFlightStatus enum with values ENABLED for active flights and STOPPED for manually halted flights), the flight date ranges (lift_measurement_flight.start_date and lift_measurement_flight.end_date) in the customer's time zone, and the target survey response collection progress (lift_measurement_flight.survey_lift_measurement.response_collection_ratio_micros, where 0 equals 0% and 1000000 equals 100%). To determine whether an ENABLED flight is in progress or completed, compare the current date in the customer's time zone against lift_measurement_flight.start_date and lift_measurement_flight.end_date, and check whether response_collection_ratio_micros has reached 1000000.

Query the LiftMeasurementConfig resource to retrieve overall study-level Brand Lift metrics, or query the following breakdown resources to analyze Brand Lift metrics across specific dimensions:

Include the segments.brand_lift_measurement_type segment to break down Brand Lift metrics by surveyed question type, defined by the BrandLiftMeasurementType enum (such as RECALL or AWARENESS). When you omit the segments.brand_lift_measurement_type segment, the API returns metrics for a single auto-selected measurement type per slice. Include the segments.date segment to break down Brand Lift metrics by date (YYYY-MM-DD). The metrics.cost_per_lifted_cookie metric is returned in micros of the study-owning account's currency.

In Google Ads API v27 and later, you can also select the read-only (OUTPUT_ONLY) lift_guidance field (lift_guidance.outcome and lift_guidance.guidance_message, of type LiftGuidance) in the Google Ads Query Language (GAQL) SELECT clause on lift_measurement_config and on the five Brand Lift breakdown resources (lift_measurement_age_range, lift_measurement_campaign, lift_measurement_device, lift_measurement_gender, and lift_measurement_video) to receive row-level statistical interpretation. On lift_measurement_config, lift_guidance is populated only when the query also selects or filters on Lift metrics or segments (it is omitted on metadata-only configuration queries), and lift_guidance.guidance_message can contain inline Markdown links (such as [Learn more](https://support.google.com/...)). You cannot filter or sort on lift_guidance subfields in WHERE or ORDER BY clauses. For descriptions of the LiftOutcomeEnum.LiftOutcome values returned in lift_guidance.outcome, see Interpret Brand Lift outcomes.

Reportability thresholds and empty responses

Brand Lift reporting is both latency-gated (requiring a 24–48 hour pipeline delay after a study starts) and response-gated per surveyed measurement type (segments.brand_lift_measurement_type) to satisfy statistical reportability and privacy thresholds:

  • Positive response rates (metrics.brand_lift_baseline_positive_response_rate* and metrics.brand_lift_exposed_positive_response_rate*): At least 50 survey responses in the corresponding arm (exposed or suppressed).
  • Strict positive lift threshold (overall study on LiftMeasurementConfig or independent slice): Statistically positive lift (p_value <= 0.50) with at least 2,000 total responses and 860 responses per arm.
  • Relaxed slice positive lift threshold (on a breakdown resource when the overall study meets the strict threshold with positive lift): Statistically positive lift (p_value <= 0.50) with at least 500 total responses and 200 responses per arm.
  • Complete non-positive lift threshold (NO_SIGNIFICANT_LIFT versus NOT_ENOUGH_DATA): At least 4,100 total responses per metric when no statistically positive lift is detected.

Queries to LiftMeasurementConfig and the five Brand Lift breakdown resources (LiftMeasurementAgeRange, LiftMeasurementGender, LiftMeasurementDevice, LiftMeasurementCampaign, and LiftMeasurementVideo) return an empty response (results: []) rather than an error in the following cases:

  • The queried customer_id parameter is not on the allowlist or doesn't own the study.
  • The WHERE date filters fall outside evaluated flight dates (lift_measurement_flight.start_date and lift_measurement_flight.end_date).
  • The study or specific demographic, device, campaign, or video slice hasn't met the 24–48 hour reporting delay and survey response volume thresholds, even when LiftMeasurementConfig returns overall study results.

To troubleshoot empty responses, complete the following steps:

  1. Query lift_measurement_flight to verify that lift_measurement_flight.status = 'ENABLED', check that your WHERE date filters fall within lift_measurement_flight.start_date and lift_measurement_flight.end_date, and inspect survey progress in lift_measurement_flight.survey_lift_measurement.response_collection_ratio_micros (where 1000000 equals 100%).
  2. Query LiftMeasurementConfig to confirm whether the overall study has generated reportable Brand Lift metrics before querying dimensional breakdown resources.

Interpret Brand Lift outcomes

In Google Ads API v27 and later, the lift_guidance.outcome field returns a LiftOutcomeEnum.LiftOutcome value that classifies the statistical state of each returned row alongside human-readable remediation advice in lift_guidance.guidance_message:

  • In-progress study outcomes (when lift_measurement_flight.status is ENABLED and the current date in the customer's time zone is on or before lift_measurement_flight.end_date with lift_measurement_flight.survey_lift_measurement.response_collection_ratio_micros less than 1000000):
    • STUDY_IN_PROGRESS_WITH_LIFT: The study is in progress and preliminary positive lift is emerging.
    • STUDY_IN_PROGRESS_NO_LIFT_YET: The study is in progress with sufficient initial volume, and no statistically significant lift has been detected yet.
    • STUDY_IN_PROGRESS_COLLECTING_DATA: The study is in progress and actively collecting initial survey responses toward statistical reporting thresholds.
  • Completed study outcomes (when lift_measurement_flight.end_date has passed or lift_measurement_flight.survey_lift_measurement.response_collection_ratio_micros has reached 1000000):
    • POSITIVE_LIFT_DETECTED: Statistically significant positive lift was detected for the completed study or slice.
    • NO_SIGNIFICANT_LIFT: The completed study or slice met statistical precision thresholds (at least 4,100 total responses per metric), and no statistically significant positive lift was detected.
    • NOT_ENOUGH_DATA: The completed study or slice did not meet the survey response thresholds required for statistical reportability (at least 2,000 total and 860 per-arm responses for positive lift on the overall study, 500 total and 200 per-arm responses for positive lift on a slice when the overall study has positive lift, or 4,100 total responses per metric when no positive lift is detected).

Apply the following guidelines when interpreting Brand Lift results:

  • In-progress versus completed flights (STUDY_IN_PROGRESS_*): While an ENABLED study flight is still in progress (before lift_measurement_flight.end_date passes and before response_collection_ratio_micros reaches 1000000), survey responses and bias-corrected modeling are still stabilizing. Treat in-progress metrics as preliminary and avoid pausing campaigns or reducing budgets until the flight completes.
  • Positive lift (POSITIVE_LIFT_DETECTED): When positive Brand Lift is detected, query the dimensional breakdown resources (LiftMeasurementAgeRange, LiftMeasurementGender, LiftMeasurementDevice, LiftMeasurementCampaign, and LiftMeasurementVideo) to identify high-performing audience cohorts ("bright spots") and evaluate creative resonance across the brand funnel.
  • Insufficient data or no significant lift (NOT_ENOUGH_DATA or NO_SIGNIFICANT_LIFT): When survey response volume is insufficient (fewer than 4,100 total responses per metric without positive lift, or fewer than the 2,000/860 strict or 500/200 relaxed slice thresholds for positive lift) or no statistically significant lift is detected on a slice, do not treat this as campaign failure or exclude audiences based on single-slice sample insufficiency. Check overall study metrics on LiftMeasurementConfig, query other surveyed segments.brand_lift_measurement_type values (such as upper-funnel RECALL or AWARENESS), and check remaining breakdown dimensions.

Example Brand Lift queries

Because flight metadata and lift statistics reside on separate resources, querying flight details and performance statistics requires separate GAQL queries.

Flight details query

To retrieve metadata for a Brand Lift flight, such as its status, measurement type, scheduled dates, and survey collection progress, query the LiftMeasurementFlight resource:

SELECT
  lift_measurement_flight.lift_measurement_config_id,
  lift_measurement_flight.lift_measurement_flight_id,
  lift_measurement_flight.name,
  lift_measurement_flight.status,
  lift_measurement_flight.lift_type,
  lift_measurement_flight.start_date,
  lift_measurement_flight.end_date,
  lift_measurement_flight.survey_lift_measurement.response_collection_ratio_micros
FROM lift_measurement_flight
WHERE lift_measurement_flight.lift_measurement_config_id = CONFIG_ID
  AND lift_measurement_flight.lift_type = 'SURVEY'

Replace the following:

  • CONFIG_ID: the integer ID of the lift measurement configuration (lift_measurement_config.lift_measurement_config_id).

Statistics query

To retrieve Brand Lift metrics, query the LiftMeasurementConfig resource or the appropriate breakdown resource. For example, the following query retrieves Brand Lift metrics segmented by campaign and Brand Lift measurement type from the LiftMeasurementCampaign resource:

SELECT
  lift_measurement_campaign.lift_measurement_config_id,
  lift_measurement_campaign.campaign,
  segments.brand_lift_measurement_type,
  metrics.absolute_brand_lift,
  metrics.relative_brand_lift
FROM lift_measurement_campaign
WHERE lift_measurement_campaign.lift_measurement_config_id = CONFIG_ID

Replace the following:

  • CONFIG_ID: the integer ID of the lift measurement configuration.

Conversion Lift reporting

In Google Ads API v25 and later, Conversion Lift reporting measures the incremental impact of your ads on conversions and conversion value. Conversion Lift studies compare converting behavior between an exposed group (users who were eligible to see your ads) and an unexposed control group (users held back from seeing your ads).

Conversion Lift configurations, flights, and experiment links are read-only in the Google Ads API (all fields are OUTPUT_ONLY) and are configured in the Google Ads UI or with your Google representative or GPS contact. Query the lift_measurement_config resource by using the customer_id parameter of the account that owns the study. A manager account can directly query cross-account Conversion Lift studies that it owns. Query studies owned by a child account by using the child account's customer_id parameter (optionally authenticated with the login-customer-id header).

Conversion Lift reporting supports two study workflows, both of which share the same study and flight discovery, cumulative snapshot date ranges, and conversion action scopes:

  • Standalone Conversion Lift reporting: Measures overall incrementality, cost per incremental conversion (iCPA), and incremental return on ad spend (iROAS) across the selected conversion actions or all conversion actions from the measured campaigns in the study.
  • A/B + Conversion Lift reporting: Measures per-arm incrementality and Bayesian winner scores between two competing experiment arms (control versus treatment) alongside overall study incrementality.

For assistance constructing valid queries to the lift_measurement_config resource, use the Google Ads Query Builder tool.

In Google Ads API v27 and later, you can select the read-only (OUTPUT_ONLY) lift_measurement_config.lift_guidance field (lift_guidance.outcome and lift_guidance.guidance_message, of type LiftGuidance) in the GAQL SELECT clause on lift_measurement_config to receive row-level statistical interpretation for standalone Conversion Lift and A/B + Conversion Lift queries. On lift_measurement_config, lift_guidance is populated only when the query also selects or filters on Lift metrics or segments (it is omitted on metadata-only configuration queries), and lift_guidance.guidance_message can contain inline Markdown links (such as [Learn more](https://support.google.com/...)). You cannot filter or sort on lift_guidance subfields in WHERE or ORDER BY clauses. The lift_guidance.outcome field returns one of the following LiftOutcomeEnum.LiftOutcome values:

  • In-progress study outcomes (when lift_measurement_flight.status is ENABLED and the current date in the customer's time zone is on or before lift_measurement_flight.end_date):
    • STUDY_IN_PROGRESS_WITH_LIFT: The study is in progress and preliminary positive lift is emerging.
    • STUDY_IN_PROGRESS_NO_LIFT_YET: The study is in progress with sufficient initial volume, and no statistically significant lift has been detected yet.
    • STUDY_IN_PROGRESS_COLLECTING_DATA: The study is in progress and actively collecting initial conversion data toward statistical reporting thresholds.
  • Completed study outcomes (when lift_measurement_flight.end_date has passed):
    • POSITIVE_LIFT_DETECTED: Statistically significant positive lift was detected for the completed study, experiment arm, or slice.
    • AB_WINNER_WITH_LIFT: An A/B + Conversion Lift study completed with a winning experiment arm head-to-head and statistically significant positive lift over the unexposed holdback group.
    • AB_WINNER_PARITY_VS_CONTROL: An A/B + Conversion Lift study completed with a winning experiment arm head-to-head, but at parity with the unexposed holdback group.
    • NO_SIGNIFICANT_LIFT: The completed study, experiment arm, or slice met statistical precision thresholds, and no statistically significant positive lift was detected.
    • NOT_ENOUGH_DATA: The completed study, experiment arm, or slice did not accumulate enough conversions to reach statistical precision thresholds.

Discover study configurations and flights

To report on a Conversion Lift study, first discover the study configuration and its flight schedule:

  1. Retrieve the study configuration: Query the LiftMeasurementConfig resource to retrieve the study's name, holdback ratio (lift_measurement_config.conversion_lift_holdback_ratio_micros), tracked conversion actions (lift_measurement_config.conversion_actions), and—in Google Ads API v26 and later—actively linked campaigns (lift_measurement_config.campaigns). The conversion_lift_holdback_ratio_micros field is expressed in micros where 1,000,000 equals 100% (for example, 300000 represents a 30% unexposed holdback group). If you don't know the study's CONFIG_ID, omit the WHERE clause to list all study configurations in the account.
  2. Retrieve flight observation windows: Query the LiftMeasurementFlight resource to find the scheduled flight dates (lift_measurement_flight.start_date and lift_measurement_flight.end_date) in the customer's time zone. Because a single LiftMeasurementConfig resource can group Brand Lift (SURVEY), Conversion Lift (CONVERSION), and Search Lift (SEARCH) flights, include the lift_measurement_flight.lift_type field in the SELECT clause and filter by lift_measurement_flight.lift_type = 'CONVERSION' in the WHERE clause. The lift_measurement_flight.start_date and lift_measurement_flight.end_date fields are sortable in ORDER BY clauses, but you cannot filter on them in a WHERE clause.

The following query retrieves the study configuration (omit the lift_measurement_config.campaigns field in Google Ads API v25, as that field requires Google Ads API v26 and later):

SELECT
  lift_measurement_config.lift_measurement_config_id,
  lift_measurement_config.name,
  lift_measurement_config.conversion_lift_holdback_ratio_micros,
  lift_measurement_config.conversion_actions,
  lift_measurement_config.campaigns
FROM lift_measurement_config
WHERE lift_measurement_config.lift_measurement_config_id = CONFIG_ID

Replace the following:

  • CONFIG_ID: the integer ID of the lift measurement configuration (lift_measurement_config.lift_measurement_config_id).

The following query retrieves the Conversion Lift flight schedule for the study:

SELECT
  lift_measurement_flight.lift_measurement_config_id,
  lift_measurement_flight.lift_measurement_flight_id,
  lift_measurement_flight.name,
  lift_measurement_flight.status,
  lift_measurement_flight.lift_type,
  lift_measurement_flight.start_date,
  lift_measurement_flight.end_date
FROM lift_measurement_flight
WHERE lift_measurement_flight.lift_measurement_config_id = CONFIG_ID
  AND lift_measurement_flight.lift_type = 'CONVERSION'

Replace the following:

  • CONFIG_ID: the integer ID of the lift measurement configuration (lift_measurement_config.lift_measurement_config_id).

In Google Ads API v26 and later, you can also discover the lift measurement configuration linked to a campaign directly from the campaign resource by querying the campaign.lift_measurement_config field. Querying this field avoids having to cross-reference the LiftMeasurementConfig resource separately:

SELECT
  campaign.id,
  campaign.lift_measurement_config
FROM campaign
WHERE campaign.id = CAMPAIGN_ID

Replace the following:

  • CAMPAIGN_ID: the integer ID of the campaign (campaign.id).

The campaign.lift_measurement_config field returns a resource name in the format customers/CUSTOMER_ID/liftMeasurementConfigs/CONFIG_ID. If a campaign has participated in multiple lift studies, this field identifies the most recently linked configuration (active or inactive). In contrast, the lift_measurement_config.campaigns field lists only actively linked campaigns.

Query cumulative snapshot date ranges

Conversion Lift results are structured as daily cumulative evaluation snapshots starting from the flight's start date:

Both segments.conversion_lift_start_date and segments.conversion_lift_end_date are INT64 fields formatted as unquoted 8-digit integers without hyphens in YYYYMMDD format (for example, 20260501). In contrast, the lift_measurement_flight.start_date and lift_measurement_flight.end_date fields return date strings in ISO 8601 YYYY-MM-DD format (for example, 2026-05-01). Unlike in Brand Lift and Search Lift queries, the segments.date segment is incompatible with all Conversion Lift metrics and segments. Selecting the segments.date segment alongside Conversion Lift segments (such as segments.conversion_lift_start_date) in the SELECT clause returns a QueryError.PROHIBITED_FIELD_COMBINATION_IN_SELECT_CLAUSE error, whereas combining segments.date with Conversion Lift metrics returns a QueryError.PROHIBITED_SEGMENT_WITH_METRIC_IN_SELECT_OR_WHERE_CLAUSE error.

The newest computed snapshot can lag behind an active flight's end date while processing data. Always query the lift_measurement_config resource for available (segments.conversion_lift_start_date, segments.conversion_lift_end_date) pairs ordered by segments.conversion_lift_end_date DESC first to identify the most recent evaluated snapshot.

The following query retrieves the available snapshot date ranges for a study:

SELECT
  segments.conversion_lift_start_date,
  segments.conversion_lift_end_date,
  segments.conversion_lift_included_conversion_action_types
FROM lift_measurement_config
WHERE lift_measurement_config.lift_measurement_config_id = CONFIG_ID
  AND segments.conversion_lift_included_conversion_action_types = 'SELECTED'
ORDER BY segments.conversion_lift_end_date DESC

Replace the following:

  • CONFIG_ID: the integer ID of the lift measurement configuration.

Filter by conversion action scope

Conversion Lift evaluates incrementality across two scopes defined by the segments.conversion_lift_included_conversion_action_types segment:

  • SELECTED: Conversion actions selected during study setup, or all biddable (primary) conversion actions if custom conversion actions were not specified.
  • ALL: All conversion actions from the campaigns measured in the study.

When you query the lift_measurement_config resource for Conversion Lift metrics without filtering on the segments.conversion_lift_included_conversion_action_types segment, the API returns two rows per slice—one for ALL and one for SELECTED. In GAQL, any non-date segment field (and any Conversion Lift snapshot date segment) filtered in the WHERE clause must also appear in the SELECT clause. Otherwise, the API returns a QueryError.EXPECTED_REFERENCED_FIELD_IN_SELECT_CLAUSE error. When you filter by segments.conversion_lift_included_conversion_action_types = 'SELECTED' in the WHERE clause, include that segment in the SELECT clause.

Standalone Conversion Lift reporting

Standalone Conversion Lift studies measure the incremental impact of your ads against an unexposed holdback group without splitting campaigns into competing treatment arms.

Query conversion lift and iCPA

Conversion Lift reports point estimates alongside 90% one-sided confidence bounds (*_p90_lower_bound and *_p90_upper_bound, which together form an 80% two-sided confidence interval) and p-values (p_value <= 0.10 corresponds to 90% one-sided statistical significance):

The API omits Conversion Lift metrics on returned rows in the following cases:

  • The inclusive snapshot window is fewer than 4 days.
  • Unscaled observed conversion volume is less than the minimum precision thresholds (fewer than 65 observed baseline conversions in the holdback group or fewer than 150 observed exposed conversions, evaluated before scaling metrics.conversion_lift_baseline_conversions).
  • Measured lift is non-positive (metrics.incremental_conversions_p_value > 0.5 for conversion metrics, or metrics.incremental_conversion_value_p_value > 0.5 for conversion value metrics).
  • In Google Ads API v27 and later, iROAS metrics (metrics.incremental_conversion_value_per_cost*) and conversion-value winner score metrics (metrics.incremental_conversion_value_winner_score and metrics.incremental_conversion_value_per_cost_winner_score) are queried on a study that was configured during setup with conversion count rather than conversion value as its primary metric. Non-cost conversion value metrics (metrics.conversion_lift_baseline_conversion_value, metrics.conversion_lift_exposed_conversion_value, metrics.incremental_conversion_value*, and metrics.relative_conversion_value_lift*) remain populated on conversion-count-primary studies whenever conversion value lift is reportable.
  • Multiple rows or studies are aggregated because you omitted the lift_measurement_config.lift_measurement_config_id, segments.conversion_lift_start_date, segments.conversion_lift_end_date, or segments.conversion_lift_included_conversion_action_types fields from the SELECT clause.

Apply the following guidelines when interpreting standalone Conversion Lift outcomes:

  • In-progress versus completed flights (STUDY_IN_PROGRESS_*): While an ENABLED study flight is still in progress (when the current date in the customer's time zone is on or before lift_measurement_flight.end_date), baseline and treatment conversion modeling and delayed conversion attribution are still stabilizing. Treat in-progress metrics as preliminary and avoid pausing campaigns or reducing budgets until the flight completes.
  • Positive lift (POSITIVE_LIFT_DETECTED): Conversion Lift metric reportability and lift_guidance.outcome = 'POSITIVE_LIFT_DETECTED' trigger whenever measured lift is positive (p_value <= 0.50, corresponding to at least 50% certainty with a positive point estimate, whereas p_value <= 0.10 indicates 90% one-sided statistical significance where *_p90_lower_bound >= 0) and minimum precision thresholds (at least 4 snapshot days, at least 65 unscaled baseline conversions, and at least 150 unscaled exposed conversions) are met. When positive lift is detected, query dimensional breakdowns (demographics, device, country, or conversion category and conversion action) to identify high-performing cohorts ("bright spots").
  • Insufficient data or no significant lift (NOT_ENOUGH_DATA or NO_SIGNIFICANT_LIFT): When data volume is insufficient (fewer than 65 unscaled baseline conversions or fewer than 150 unscaled exposed conversions) or no statistically significant lift is detected on a slice, do not treat this as campaign failure or exclude audiences based on single-slice sample insufficiency. Check overall study metrics and query segments.conversion_lift_included_conversion_action_types = 'ALL' (if SELECTED shows no lift) or remaining dimensional breakdowns to verify whether positive lift was isolated elsewhere.

The following query retrieves overall conversion incrementality, confidence bounds, and iCPA:

SELECT
  lift_measurement_config.lift_measurement_config_id,
  segments.conversion_lift_start_date,
  segments.conversion_lift_end_date,
  segments.conversion_lift_included_conversion_action_types,
  metrics.conversion_lift_baseline_conversions,
  metrics.conversion_lift_exposed_conversions,
  metrics.incremental_conversions,
  metrics.incremental_conversions_p90_lower_bound,
  metrics.incremental_conversions_p90_upper_bound,
  metrics.incremental_conversions_p_value,
  metrics.cost_per_incremental_conversion,
  metrics.cost_per_incremental_conversion_p90_lower_bound,
  metrics.cost_per_incremental_conversion_p90_upper_bound,
  metrics.relative_conversion_lift,
  metrics.relative_conversion_lift_p90_lower_bound,
  metrics.relative_conversion_lift_p90_upper_bound
FROM lift_measurement_config
WHERE lift_measurement_config.lift_measurement_config_id = CONFIG_ID
  AND segments.conversion_lift_start_date = START_DATE
  AND segments.conversion_lift_end_date = END_DATE
  AND segments.conversion_lift_included_conversion_action_types = 'SELECTED'

Replace the following:

  • CONFIG_ID: the integer ID of the lift measurement configuration.
  • START_DATE: the snapshot start date as an unquoted 8-digit YYYYMMDD integer (for example, 20260501).
  • END_DATE: the snapshot end date as an unquoted 8-digit YYYYMMDD integer (for example, 20260515).

Query conversion value lift and iROAS

To evaluate incremental revenue generated by your ad spend, query conversion value metrics and the metrics.incremental_conversion_value_per_cost metric. Non-cost conversion value metrics (metrics.conversion_lift_baseline_conversion_value, metrics.conversion_lift_exposed_conversion_value, metrics.incremental_conversion_value*, and metrics.relative_conversion_value_lift*) are populated whenever the study has reportable conversion value lift, regardless of whether the study's primary metric is conversion count or conversion value. In Google Ads API v27 and later, the iROAS metrics (metrics.incremental_conversion_value_per_cost*)—along with metrics.incremental_conversion_value_winner_score and metrics.incremental_conversion_value_per_cost_winner_score—are returned only when the study was configured during setup with conversion value as its primary metric (and are omitted on conversion-count-primary studies):

The following query retrieves conversion value incrementality, confidence bounds, and iROAS:

SELECT
  lift_measurement_config.lift_measurement_config_id,
  segments.conversion_lift_start_date,
  segments.conversion_lift_end_date,
  segments.conversion_lift_included_conversion_action_types,
  metrics.conversion_lift_baseline_conversion_value,
  metrics.conversion_lift_exposed_conversion_value,
  metrics.incremental_conversion_value,
  metrics.incremental_conversion_value_p90_lower_bound,
  metrics.incremental_conversion_value_p90_upper_bound,
  metrics.incremental_conversion_value_p_value,
  metrics.incremental_conversion_value_per_cost,
  metrics.incremental_conversion_value_per_cost_p90_lower_bound,
  metrics.incremental_conversion_value_per_cost_p90_upper_bound,
  metrics.relative_conversion_value_lift,
  metrics.relative_conversion_value_lift_p90_lower_bound,
  metrics.relative_conversion_value_lift_p90_upper_bound
FROM lift_measurement_config
WHERE lift_measurement_config.lift_measurement_config_id = CONFIG_ID
  AND segments.conversion_lift_start_date = START_DATE
  AND segments.conversion_lift_end_date = END_DATE
  AND segments.conversion_lift_included_conversion_action_types = 'SELECTED'

Replace the following:

  • CONFIG_ID: the integer ID of the lift measurement configuration.
  • START_DATE: the snapshot start date as an unquoted 8-digit YYYYMMDD integer (for example, 20260501).
  • END_DATE: the snapshot end date as an unquoted 8-digit YYYYMMDD integer (for example, 20260515).

Segment results by dimension

Segment incremental conversion metrics by conversion category, conversion action, demographics, device, or geography. When querying the lift_measurement_config resource, select only compatible segment combinations:

Selecting mutually incompatible segments in the SELECT clause fails query validation and returns a QueryError.PROHIBITED_FIELD_COMBINATION_IN_SELECT_CLAUSE error. Use the Google Ads Query Builder to validate segment combinations interactively.

Evaluate cost-efficiency metrics (metrics.cost_per_incremental_conversion* and metrics.incremental_conversion_value_per_cost*) only on overall study rows or segments.experiment_arm rows. In Google Ads API v25 and v26, the schema permits selecting the six cost-ratio metrics (metrics.cost_per_incremental_conversion, metrics.cost_per_incremental_conversion_p90_lower_bound, metrics.cost_per_incremental_conversion_p90_upper_bound, metrics.incremental_conversion_value_per_cost, metrics.incremental_conversion_value_per_cost_p90_lower_bound, and metrics.incremental_conversion_value_per_cost_p90_upper_bound) with the eight non-arm slice segments (segments.conversion_action, segments.conversion_action_name, segments.conversion_lift_conversion_category, segments.age_range, segments.gender, segments.device, segments.country, and segments.country_localized_name). Starting in Google Ads API v27, selecting any of these six cost-ratio metrics with any of those eight non-arm slice segments fails query validation and returns a QueryError.PROHIBITED_SEGMENT_WITH_METRIC_IN_SELECT_OR_WHERE_CLAUSE error. Non-cost conversion value metrics (metrics.conversion_lift_baseline_conversion_value, metrics.conversion_lift_exposed_conversion_value, metrics.incremental_conversion_value*, and metrics.relative_conversion_value_lift*) remain supported on conversion category, conversion action, demographic, device, and country segments whenever conversion value lift is reportable (regardless of whether the study's primary metric is conversion count or conversion value).

To evaluate lift across conversion categories defined in the ConversionActionCategory enum (such as PURCHASE, SIGNUP, or SUBMIT_LEAD_FORM), segment results by the segments.conversion_lift_conversion_category segment:

SELECT
  lift_measurement_config.lift_measurement_config_id,
  segments.conversion_lift_start_date,
  segments.conversion_lift_end_date,
  segments.conversion_lift_included_conversion_action_types,
  segments.conversion_lift_conversion_category,
  metrics.conversion_lift_baseline_conversions,
  metrics.conversion_lift_exposed_conversions,
  metrics.incremental_conversions,
  metrics.relative_conversion_lift
FROM lift_measurement_config
WHERE lift_measurement_config.lift_measurement_config_id = CONFIG_ID
  AND segments.conversion_lift_start_date = START_DATE
  AND segments.conversion_lift_end_date = END_DATE
  AND segments.conversion_lift_included_conversion_action_types = 'SELECTED'

Replace the following:

  • CONFIG_ID: the integer ID of the lift measurement configuration.
  • START_DATE: the snapshot start date as an unquoted 8-digit YYYYMMDD integer (for example, 20260501).
  • END_DATE: the snapshot end date as an unquoted 8-digit YYYYMMDD integer (for example, 20260515).

The segments.conversion_action segment returns the conversion action resource name in the format customers/CUSTOMER_ID/conversionActions/CONVERSION_ACTION_ID. Select the segments.conversion_action_name segment on the lift_measurement_config resource to retrieve human-readable conversion action names without querying the conversion_action resource separately. The following query segments Conversion Lift metrics by conversion action, conversion action name, and conversion category:

SELECT
  lift_measurement_config.lift_measurement_config_id,
  segments.conversion_lift_start_date,
  segments.conversion_lift_end_date,
  segments.conversion_lift_included_conversion_action_types,
  segments.conversion_action,
  segments.conversion_action_name,
  segments.conversion_lift_conversion_category,
  metrics.conversion_lift_baseline_conversions,
  metrics.conversion_lift_exposed_conversions,
  metrics.incremental_conversions,
  metrics.relative_conversion_lift
FROM lift_measurement_config
WHERE lift_measurement_config.lift_measurement_config_id = CONFIG_ID
  AND segments.conversion_lift_start_date = START_DATE
  AND segments.conversion_lift_end_date = END_DATE
  AND segments.conversion_lift_included_conversion_action_types = 'SELECTED'

Replace the following:

  • CONFIG_ID: the integer ID of the lift measurement configuration.
  • START_DATE: the snapshot start date as an unquoted 8-digit YYYYMMDD integer (for example, 20260501).
  • END_DATE: the snapshot end date as an unquoted 8-digit YYYYMMDD integer (for example, 20260515).

The following query segments Conversion Lift metrics by the segments.device segment. Substitute the segments.age_range or segments.gender segment in separate requests to retrieve demographic breakdowns (optionally combined with conversion category or conversion action segments):

SELECT
  lift_measurement_config.lift_measurement_config_id,
  segments.conversion_lift_start_date,
  segments.conversion_lift_end_date,
  segments.conversion_lift_included_conversion_action_types,
  segments.device,
  metrics.conversion_lift_baseline_conversions,
  metrics.conversion_lift_exposed_conversions,
  metrics.incremental_conversions,
  metrics.relative_conversion_lift
FROM lift_measurement_config
WHERE lift_measurement_config.lift_measurement_config_id = CONFIG_ID
  AND segments.conversion_lift_start_date = START_DATE
  AND segments.conversion_lift_end_date = END_DATE
  AND segments.conversion_lift_included_conversion_action_types = 'SELECTED'

Replace the following:

  • CONFIG_ID: the integer ID of the lift measurement configuration.
  • START_DATE: the snapshot start date as an unquoted 8-digit YYYYMMDD integer (for example, 20260501).
  • END_DATE: the snapshot end date as an unquoted 8-digit YYYYMMDD integer (for example, 20260515).

To break down results by geography, include the segments.country and segments.country_localized_name segments (optionally combined with conversion category or conversion action segments). The segments.country segment returns a global geo target constant resource name in the format geoTargetConstants/CRITERION_ID (for example, geoTargetConstants/2840 for the United States).

The following query segments Conversion Lift metrics by country criterion resource name and localized country name:

SELECT
  lift_measurement_config.lift_measurement_config_id,
  segments.conversion_lift_start_date,
  segments.conversion_lift_end_date,
  segments.conversion_lift_included_conversion_action_types,
  segments.country,
  segments.country_localized_name,
  metrics.conversion_lift_baseline_conversions,
  metrics.conversion_lift_exposed_conversions,
  metrics.incremental_conversions,
  metrics.relative_conversion_lift
FROM lift_measurement_config
WHERE lift_measurement_config.lift_measurement_config_id = CONFIG_ID
  AND segments.conversion_lift_start_date = START_DATE
  AND segments.conversion_lift_end_date = END_DATE
  AND segments.conversion_lift_included_conversion_action_types = 'SELECTED'

Replace the following:

  • CONFIG_ID: the integer ID of the lift measurement configuration.
  • START_DATE: the snapshot start date as an unquoted 8-digit YYYYMMDD integer (for example, 20260501).
  • END_DATE: the snapshot end date as an unquoted 8-digit YYYYMMDD integer (for example, 20260515).

A/B + Conversion Lift reporting

In Google Ads API v25 and later, A/B + Conversion Lift studies combine a two-arm experiment (comparing control versus treatment strategies) with holdback measurement in the Google Ads API to assess individual arm incrementality and compute Bayesian winner scores. Unlike direct experiment reporting on the experiment resource, which compares unadjusted arm-versus-arm traffic metrics, querying the lift_measurement_config resource with the segments.experiment_arm segment measures each arm's incremental lift against an unexposed holdback baseline.

Discover experiments and arms

An A/B + Conversion Lift study links a LiftMeasurementConfig resource to a two-arm experiment resource through the read-only (OUTPUT_ONLY) experiment.lift_measurement_config field.

To discover the linked experiment and its arms in a single request, query the ExperimentArm resource with attributed experiment fields and filter on the experiment.lift_measurement_config field. The response includes each arm's control status (experiment_arm.control), traffic split (experiment_arm.traffic_split), linked campaigns (experiment_arm.campaigns), and resource name. Use the returned experiment_arm.resource_name value (formatted as customers/CUSTOMER_ID/experimentArms/EXPERIMENT_ID~ARM_ID) to filter on the segments.experiment_arm segment:

SELECT
  experiment.experiment_id,
  experiment.name,
  experiment.lift_measurement_config,
  experiment_arm.resource_name,
  experiment_arm.name,
  experiment_arm.control,
  experiment_arm.traffic_split,
  experiment_arm.campaigns
FROM experiment_arm
WHERE experiment.lift_measurement_config =
  'customers/CUSTOMER_ID/liftMeasurementConfigs/CONFIG_ID'

Replace the following:

  • CUSTOMER_ID: the 10-digit customer ID of the account that owns the study (without hyphens).
  • CONFIG_ID: the integer ID of the lift measurement configuration.

If you start from a known EXPERIMENT_ID (or want to list all lift-enabled experiments in the account), run the same experiment_arm query filtered by WHERE experiment.experiment_id = EXPERIMENT_ID (or WHERE experiment.lift_measurement_config IS NOT NULL).

Query per-arm incrementality and winner scores

To measure incrementality for each arm and evaluate the statistical probability that an arm is the winner, include the segments.experiment_arm segment in your query to the lift_measurement_config resource.

The following Bayesian winner score metrics evaluate the probability (from 0.0 to 1.0, summing to 1.0 across both arms when both arms have reportable positive lift) that a given arm is superior:

  • metrics.incremental_conversions_winner_score: The probability that this arm is the winner for incremental conversions, normalized by arm traffic split (experiment_arm.traffic_split).
  • metrics.incremental_conversion_value_winner_score: The probability that this arm is the winner for incremental conversion value, normalized by arm traffic split (experiment_arm.traffic_split). In Google Ads API v27 and later, populated only when the study's primary metric is conversion value, and omitted (left unset) on conversion-count-primary studies.
  • metrics.cost_per_incremental_conversion_winner_score: The probability that this arm is the winner for cost per incremental conversion (lowest iCPA), normalized by spend.
  • metrics.incremental_conversion_value_per_cost_winner_score: The probability that this arm is the winner for incremental conversion value per cost (highest iROAS), normalized by spend. In Google Ads API v27 and later, populated only when the study's primary metric is conversion value, and omitted (left unset) on conversion-count-primary studies.

The following query retrieves per-arm incrementality and Bayesian winner scores for both arms:

SELECT
  lift_measurement_config.lift_measurement_config_id,
  segments.conversion_lift_start_date,
  segments.conversion_lift_end_date,
  segments.conversion_lift_included_conversion_action_types,
  segments.experiment_arm,
  metrics.conversion_lift_baseline_conversions,
  metrics.conversion_lift_exposed_conversions,
  metrics.incremental_conversions,
  metrics.cost_per_incremental_conversion,
  metrics.incremental_conversions_winner_score,
  metrics.cost_per_incremental_conversion_winner_score,
  metrics.incremental_conversion_value,
  metrics.incremental_conversion_value_winner_score,
  metrics.incremental_conversion_value_per_cost_winner_score
FROM lift_measurement_config
WHERE lift_measurement_config.lift_measurement_config_id = CONFIG_ID
  AND segments.conversion_lift_start_date = START_DATE
  AND segments.conversion_lift_end_date = END_DATE
  AND segments.conversion_lift_included_conversion_action_types = 'SELECTED'

Replace the following:

  • CONFIG_ID: the integer ID of the lift measurement configuration.
  • START_DATE: the snapshot start date as an unquoted 8-digit YYYYMMDD integer (for example, 20260501).
  • END_DATE: the snapshot end date as an unquoted 8-digit YYYYMMDD integer (for example, 20260515).

To filter for a single arm, add AND segments.experiment_arm = 'customers/CUSTOMER_ID/experimentArms/EXPERIMENT_ID~ARM_ID' to the WHERE clause, where EXPERIMENT_ID is the experiment ID and ARM_ID is the arm ID returned in the experiment_arm.resource_name field. The segments.experiment_arm segment supports filtering in WHERE clauses (=, !=, IN, NOT IN), but you cannot sort on it in an ORDER BY clause.

Interpret A/B + Conversion Lift outcomes

An A/B + Conversion Lift study evaluates two hypotheses simultaneously: head-to-head arm superiority (metrics.*_winner_score) and causal lift against the unexposed holdback group (metrics.incremental_* and *_p_value). While an ENABLED study flight is still in progress (STUDY_IN_PROGRESS_*, when the current date in the customer's time zone is on or before lift_measurement_flight.end_date), head-to-head arm comparison and holdback conversion modeling are still stabilizing. Treat in-progress metrics as preliminary and avoid promoting an experiment arm or reallocating budgets until the flight completes.

On completed studies, an experiment arm is classified as a head-to-head winner when it meets the minimum precision thresholds (at least 4 snapshot days, 65 unscaled baseline conversions, and 150 unscaled exposed conversions), has 0 <= p_value <= 0.50, and achieves a Bayesian winner score greater than 0.55 (> 0.55) on the study's primary metric. This dual-hypothesis evaluation produces two distinct winner states (also reflected in lift_guidance.outcome in Google Ads API v27 and later):

  • Clear winner with causal lift (AB_WINNER_WITH_LIFT): The winning arm outperforms the competing arm head-to-head (winner_score > 0.55) and generates statistically significant incremental conversions or conversion value over the unexposed holdback group. Promote or scale the winning arm's strategy, and query dimensional breakdowns to review high-performing cohorts.
  • Head-to-head winner at parity with holdback (AB_WINNER_PARITY_VS_CONTROL): The winning arm outperforms the competing arm head-to-head (winner_score > 0.55), but neither arm produces statistically significant incremental lift over the unexposed holdback group at the queried level. Before scaling spend on the winning arm, query ALL conversions (segments.conversion_lift_included_conversion_action_types = 'ALL') and segments.conversion_lift_conversion_category breakdowns to verify whether positive causal lift was isolated elsewhere.

The POSITIVE_LIFT_DETECTED outcome applies across A/B + Conversion Lift query granularities as follows:

  • Combined-arm (OVERALL) study rows (queries without segments.experiment_arm): The study generated positive lift over the unexposed holdback group (p_value <= 0.50), and neither arm achieved a head-to-head winner score greater than 0.55.
  • Per-arm (segments.experiment_arm) rows: When one arm wins (winner_score > 0.55) and both arms achieve positive lift over the holdback group, the winning arm row returns AB_WINNER_WITH_LIFT while the non-winning arm row returns POSITIVE_LIFT_DETECTED.
  • Conversion category (segments.conversion_lift_conversion_category) slice rows: Positive lift rows always return POSITIVE_LIFT_DETECTED because winner scores are not evaluated per conversion category.

When a slice returns NOT_ENOUGH_DATA or NO_SIGNIFICANT_LIFT, do not pause arms or exclude audiences based on single-slice sample insufficiency. Check overall study metrics and ALL conversions (segments.conversion_lift_included_conversion_action_types = 'ALL').

Query combined-arm and multi-study results

When you omit the segments.experiment_arm segment from a query on the lift_measurement_config resource, the API returns study-level incrementality aggregated across both combined arms. On combined-arm queries, omit the four metrics.*_winner_score metrics, which apply only across competing arms and are omitted (left unset) on combined-arm rows without segments.experiment_arm. Use the study-level queries in Query conversion lift and iCPA and Query conversion value lift and iROAS.

When you segment A/B + Conversion Lift results by the segments.experiment_arm segment, the API supports only the segments.conversion_lift_conversion_category segment as an additional breakdown dimension. Because arm spend is not attributed to individual conversion categories, evaluate conversion and conversion value lift metrics (omitting iCPA and iROAS metrics, which fail query validation with a QueryError.PROHIBITED_SEGMENT_WITH_METRIC_IN_SELECT_OR_WHERE_CLAUSE error in Google Ads API v27 and later) when you include the segments.conversion_lift_conversion_category segment. The segments.conversion_action, segments.conversion_action_name, segments.age_range, segments.gender, segments.device, segments.country, segments.country_localized_name, and segments.date segments are incompatible with the segments.experiment_arm segment and with the metrics.*_winner_score metrics. Selecting any of these incompatible segments alongside segments.experiment_arm returns a QueryError.PROHIBITED_FIELD_COMBINATION_IN_SELECT_CLAUSE error (and a QueryError.PROHIBITED_SEGMENT_WITH_METRIC_IN_SELECT_OR_WHERE_CLAUSE error if the query also selects any metrics.*_winner_score metric).

To retrieve results for multiple study configurations in a single request, filter with the IN operator:

SELECT
  lift_measurement_config.lift_measurement_config_id,
  segments.conversion_lift_start_date,
  segments.conversion_lift_end_date,
  segments.conversion_lift_included_conversion_action_types,
  segments.experiment_arm,
  segments.conversion_lift_conversion_category,
  metrics.incremental_conversions,
  metrics.incremental_conversions_winner_score,
  metrics.incremental_conversion_value,
  metrics.incremental_conversion_value_winner_score
FROM lift_measurement_config
WHERE lift_measurement_config.lift_measurement_config_id IN (
    CONFIG_ID_1, CONFIG_ID_2)
  AND segments.conversion_lift_start_date = START_DATE
  AND segments.conversion_lift_end_date = END_DATE
  AND segments.conversion_lift_included_conversion_action_types = 'SELECTED'

Replace the following:

  • CONFIG_ID_1: the integer ID of the first lift measurement configuration.
  • CONFIG_ID_2: the integer ID of the second lift measurement configuration.
  • START_DATE: the snapshot start date as an unquoted 8-digit YYYYMMDD integer (for example, 20260501).
  • END_DATE: the snapshot end date as an unquoted 8-digit YYYYMMDD integer (for example, 20260515).

Exact date equality (= START_DATE and = END_DATE) returns rows only for studies that share that exact snapshot window. When querying studies with different flight start dates or latest evaluated snapshot dates, filter with range operators (>= and <=) or omit the date filters from the WHERE clause while keeping both the segments.conversion_lift_start_date and segments.conversion_lift_end_date segments in the SELECT clause.

Search Lift reporting

To measure the increase in organic search query volume driven by your ad campaigns in Google Ads API v26 and later, query Search Lift reporting resources. Search Lift quantifies the difference in search behavior for your brand or products between users exposed to your ads and an unexposed control group.

Search Lift resources and segmentation

Query the SearchLiftKeywordBundle resource to retrieve details about keyword bundles configured for Search Lift studies. The SearchLiftKeywordBundle resource is read-only in the Google Ads API (all fields are OUTPUT_ONLY), and keyword bundles are configured in the Google Ads UI or by your Google representative or GPS contact. Its queryable fields in GAQL are resource_name, search_lift_keyword_bundle_id, source (SearchLiftMatchingInfoSource), and status (SearchLiftKeywordBundleStatus) (the repeated keywords field is not selectable in GAQL queries).

Query the LiftMeasurementConfig resource to retrieve Search Lift metrics and overall study configurations.

In Google Ads API v27 and later, you can also select the read-only (OUTPUT_ONLY) lift_measurement_config.lift_guidance field (lift_guidance.outcome and lift_guidance.guidance_message, of type LiftGuidance) in the GAQL SELECT clause on lift_measurement_config when querying Search Lift metrics or segments (segments.search_lift_keyword_bundle or segments.search_property). On lift_measurement_config, lift_guidance is omitted on metadata-only configuration queries, and lift_guidance.guidance_message can contain inline Markdown links (such as [Learn more](https://support.google.com/...)). For Search Lift, lift_guidance.outcome returns STUDY_IN_PROGRESS_WITH_LIFT, STUDY_IN_PROGRESS_NO_LIFT_YET, STUDY_IN_PROGRESS_COLLECTING_DATA, POSITIVE_LIFT_DETECTED (at least 70% certainty, p_value <= 0.30), NO_SIGNIFICANT_LIFT, or NOT_ENOUGH_DATA (evaluated against thresholds of at least 1,000,000 overall impressions or 100,000 slice impressions, at least 75 observed searches, and at least 17 of 20 cookie buckets, where a single underpowered keyword bundle causes overall study results to report NOT_ENOUGH_DATA).

Segment Search Lift performance across the following dimensions:

  • segments.search_lift_keyword_bundle: Segment by the keyword bundle resource name. Compatible with the segments.search_property, segments.date, and demographic segments (segments.device, segments.age_range, and segments.gender). Every keyword bundle reports the full campaign cost and impressions, and keywords can belong to more than one bundle. Do not sum metrics.search_lift_cost_micros, metrics.search_lift_impressions, or Search Lift metrics across keyword bundles.
  • segments.search_property: Segment by the search property where the search occurred, defined by the SearchLiftSearchProperty enum (such as GOOGLE_SEARCH or YOUTUBE). The segments.search_property segment is compatible only with the segments.search_lift_keyword_bundle and segments.date segments. Combining segments.search_property with demographic segments (segments.device, segments.age_range, or segments.gender) fails query validation. Because segments.search_property partitions where subsequent organic searches occurred rather than where ads were served, rows segmented by segments.search_property repeat unpartitioned campaign-level cost and impressions in Google Ads API v26. Starting in Google Ads API v27, all eight spend and impression metrics (metrics.search_lift_cost_micros, metrics.search_lift_impressions, metrics.normalized_incremental_searches_per_cost*, and metrics.normalized_incremental_searches_per_impression*) are incompatible with segments.search_property and fail query validation with a QueryError.PROHIBITED_SEGMENT_WITH_METRIC_IN_SELECT_OR_WHERE_CLAUSE error.
  • Demographic and date segments: segments.device, segments.age_range, segments.gender, and segments.date.

Search Lift metrics

The LiftMeasurementConfig resource provides Search Lift metrics to evaluate search behavior changes:

The following normalized incremental metrics—along with their corresponding 90% one-sided confidence bounds (*_p90_lower_bound and *_p90_upper_bound)—are reported when lift is statistically positive:

Example Search Lift queries

The following GAQL queries demonstrate how to retrieve Search Lift keyword bundles and performance metrics.

Keyword bundle query

To retrieve keyword bundles configured for Search Lift measurements, query the search_lift_keyword_bundle resource:

SELECT
  search_lift_keyword_bundle.resource_name,
  search_lift_keyword_bundle.search_lift_keyword_bundle_id,
  search_lift_keyword_bundle.source,
  search_lift_keyword_bundle.status
FROM search_lift_keyword_bundle

Metrics query

The search_lift_keyword_bundle.resource_name values retrieved in the keyword bundle query correspond to the values returned in the segments.search_lift_keyword_bundle segment in the following query. To discover the CONFIG_ID, linked campaigns, and flight schedule for your study, follow Discover study configurations and flights on the lift_measurement_config and lift_measurement_flight resources (filtering flights by lift_measurement_flight.lift_type = 'SEARCH').

The following query retrieves Search Lift metrics segmented by keyword bundle and search property (omit segments.search_property when selecting metrics.search_lift_cost_micros or metrics.search_lift_impressions):

SELECT
  lift_measurement_config.lift_measurement_config_id,
  segments.search_lift_keyword_bundle,
  segments.search_property,
  metrics.relative_search_lift,
  metrics.relative_search_lift_p90_lower_bound,
  metrics.relative_search_lift_p90_upper_bound,
  metrics.incremental_searches_p_value,
  metrics.normalized_incremental_searches,
  metrics.normalized_incremental_searches_p90_lower_bound,
  metrics.normalized_incremental_searches_p90_upper_bound
FROM lift_measurement_config
WHERE lift_measurement_config.lift_measurement_config_id = CONFIG_ID

Replace the following:

  • CONFIG_ID: the integer ID of the lift measurement configuration.
Previous
Reporting
Next
Pull geo data