Brand Lift Studies in Google Ads: Setup, Budgets and Metrics

Everything here comes from Google Ads Help: About Brand Lift, Set up Brand Lift, Understand Lift measurement statuses and metrics and About certainty of lift. Checked on 12 November 2026. For testing campaign settings against each other, see our Google Ads experiments guide; for the video formats themselves, see YouTube video ads.

What Brand Lift measures

Google describes Brand Lift as a free tool for measuring the effectiveness of your video ads in Video and Demand Gen campaigns. The setup page also lists audio campaigns. Instead of clicks, impressions or views, it measures how people feel about your brand or product. You can pick up to three of these metrics per study:

• Ad recall (the question wording adapts: "hearing" for audio-only ads, "viewing" for video-only, "encountering" for both)
• Awareness
• Association
• Consideration
• Favorability
• Purchase intent
• A custom metric

For Demand Gen, Google notes that measurement won't capture results from other Demand Gen channels like Gmail or Discover, so it advises prioritizing video ads on YouTube in those campaigns.

Who can use it

Brand Lift is not open to every account. Both help pages say you need a Google account representative: Without a representative, you won't be able to use this feature. If you do not see the Lift studies option, that is the most likely reason.

How the surveys work

Once your campaigns run, Google shows short survey questions on YouTube, before a video starts, to two groups:

• the exposed group: people who saw your ads;
• the baseline (control) group: people who were eligible to see your ads but were not shown them.

The difference in positive answers between the two groups is the lift. A viewer gets at most one survey a day and three a week, and can answer once per product or brand in 14 days. If you measure several metrics, each viewer sees only one of those questions, chosen at random. You only get aggregated answers; respondents are not identified.

Setting up a study

1. Go to Experiments in the Campaigns menu, open the Lift studies tab and select the plus button.
2. Under "Select Lift types", tick Brand lift.
3. Name the study after the product or brand. Google asks for a specific name ("Google Pixel 4" rather than "Pixel") that matches the preferred answer in the survey.
4. Select the campaigns to measure. Demand Gen campaigns must run on channels that support Brand Lift.
5. Under "Study settings", choose the strength of the study (see Standard vs Enhanced Lift).
6. Pick the survey language. It should match the language of the ad; for ads in several languages, Google suggests one product or brand per language.
7. Choose up to three metrics, then the question wording through "Product or brand type" (for example Automotive - Product) and "Final intended action" (for example Buy).
8. Enter your product or brand as the preferred answer and up to three competitor names, check the preview and save.

Surveys take up to 48 hours to be approved and start automatically when the campaigns run. The "Measurement eligibility" box on the right shows whether the selected campaigns meet the minimum budget. Paused campaigns or campaigns with a future start date do not count towards it until they run.

Two setup traps from the same page: you cannot add an audio campaign to a video-only Ad recall study or the reverse, so include both types at creation if you plan to run both. And for YouTube Reservation campaigns the study has to be set up in the campaign itself (Settings > Edit campaign > Lift Measurement); adding it later from the study gives the error This campaign can't be selected because it's an instant reserve campaign.

Minimum budgets

The minimum is set per product or brand, summed across all active campaigns in the study, and has to be met within 10 days of the study start. It depends on the number of questions and the targeted country:

Country B includes the United States, United Kingdom, France, Spain, Ireland, Sweden, New Zealand, Israel, Taiwan, Greece and Croatia. Country C includes Australia, Canada, Germany, Italy, Japan, the Netherlands, Denmark, Finland, Norway, Austria, Belgium, Switzerland, South Korea and Qatar. Country A covers the long list of other countries on the setup page, such as Brazil, India, Mexico, Poland and Turkey.

• Targeting countries in different groups uses the highest group. Targeting a country not on the lists, or no specific country, also uses the highest group.
• Amounts are in USD and converted to your account currency with exchange rates refreshed four times a year.
• The calculator looks at campaign budgets and dates, not spend to date. A campaign running longer than 10 days counts as daily budget x 10; shorter campaigns count only the days inside the study.
• With a campaign total budget, Google divides it by the campaign's days. Extending the end date lowers that daily figure and can make the study ineligible.

Meeting the minimum increases but doesn't guarantee results; reach matters too. How campaign budgets work in general is covered in Google Ads budgets and bidding.

Standard vs Enhanced Lift

Google suggests Enhanced Lift when you want higher confidence, or when earlier studies for similar campaigns ended with "No lift" or "Not enough data" despite large spend. It improves the chances of detecting lift but does not guarantee it. Cutting the budget below the threshold mid-study pauses the study.

Survey rules that get studies rejected

• No personal or sensitive data: studies cannot collect personally identifiable information or ask about topics such as health, religion, political affiliation, sexual orientation, ethnicity, financial status or children under 13.
• Spelling: brand and product names must match official spelling (logos, packaging, the official website). Misspellings are a common rejection reason.
• Language: questions and answers must match the survey language.
• One opt-out: multiple-answer questions already add "None of the above", so a manual "None" or "Not applicable" is flagged.
• Relevant answers: answer options must fit the product type and intended action you picked; random letters, numbers or emoji are not allowed.

Repeated or serious violations can suspend the survey or the whole Google Ads account.

How long results take

Spend has to be met in the first 10 days. Studies then usually finish within 14 days, sometimes earlier: a study often completes when it reaches its target of roughly 4,100 responses per metric, and Google says ending early this way does not make the results less valid. A study generally cannot be extended; you add a new one (re-measurement) instead.

Google's guide to response counts:

• about 2,000 responses per metric to detect lift in high-performing campaigns;
• 4,100 responses per metric at the recommended budget minimum;
• no lift after 16,800 responses per metric means you may not be able to detect it.

The smaller the lift, the more responses it takes:

Below 50% progress you see "Not enough data". Between 50% and 100%, significant positive lift is reported, otherwise "Not enough data". At 100%, it is lift or "No lift detected". Response counts update once a day, and results can move during the study, so Google recommends waiting for the final report.

Reading the metrics

Google's example for the three lift percentages uses positive answers rising from 20% in the baseline group to 40% in the exposed group:

• Absolute Brand Lift: exposed rate minus baseline rate. In the example, 20%.
• Relative Brand Lift: that difference divided by the baseline rate. In the example, 100%.
• Headroom Brand Lift: absolute lift divided by (1 minus the baseline rate), the share of the possible gain you achieved. In the example, 25%.
• Lifted users: the estimated number of people whose answer changed because of your ads, scaled to the campaign's reach. Not necessarily unique users. A co-viewed version adds people watching together on connected TVs.
• Cost per lifted user: campaign cost divided by lifted users.
• Baseline and exposed positive response rates, and the response counts for each group.

Google recommends cost per lifted user as the main success metric, because it combines reach and cost. Its example: two campaigns with the same cost, one with 10% absolute lift at a $15 CPM and one with 5% at a $5 CPM. The second reached more people and had 50% more lifted users at a 33% lower cost per lifted user.

Confidence and certainty. The metrics page describes lift ranges as 80% two-sided confidence intervals, meaning a 90% chance the true lift is above the lower bound (one paragraph on the same page says "usually 90%", so read the range as Google's 80% interval). Certainty of lift is 1 minus the p-value. Google reports results from 50% certainty upwards: 90% or more means the result is very unlikely to be chance, 70-90% a good chance, 50-70% directional only. Below 50% shows as "No lift", which Google says does not necessarily mean the ads did nothing.

Where to find results

Brand Lift columns are available in most tables (Campaign, Ad group, Demographics and more): open the columns icon, Modify columns, then Brand Lift. To split by metric, use the segment icon and Brand lift type. Product or brand level results are in the Lift studies table. The help pages place Lift studies in two menus: setup starts from Experiments in the Campaigns menu, while the results and re-measurement steps link to Lift studies in the Goals menu.

Re-measurement: once a study is complete, add a new study under the same product or brand to measure again, for example after new creatives or a major media change. Google also emails users with account access when a study completes or becomes ineligible, unless they turned off performance reporting emails.

"Not enough data" and "No lift detected"

For Not enough data, Google's checklist:

• Spend the budget in full; actual spend has to meet the minimum, not just the budget.
• A CPV bid that is too low can lose the auction and limit traffic.
• Targeting people who already viewed the ad (remarketing lists, later steps of a video ad sequence) means no clean control group can be built.
• Narrow targeting (placements, keywords, retargeting, small areas) cuts the number of viewers who can be surveyed. Country level is ideal.
• Target the survey's language only; surveys shown to people who do not speak it get dismissed.
• Many campaigns or experiment arms in one study thin out responses per campaign.
• For reach goals, a frequency cap spreads the ads across more people (see frequency capping), though Google notes higher frequency usually gives more lift.

For No lift detected: check segments (age, gender, device, campaign) for lift; pick realistic competitor answers, since a small brand listed next to a global leader loses on recognition; make sure your brand is the preferred answer; measure the product, not the whole brand, when the ad is about one product; show the brand early in the ad; and avoid other campaigns, channels or studies showing the same creative, which contaminates the control group.

Our view

This is our own opinion, not Google's guidance. Brand Lift only makes sense when a campaign's job is awareness and the budget already clears the minimum for your country. If you would have to raise spend just to qualify, a geo split or a simpler before-and-after check of branded search is usually the better first step. When we plan a study, we pick one or two metrics instead of three (the minimum doubles or triples), keep targeting broad and in one language, and judge the result on cost per lifted user rather than the lift percentage alone.

Frequently asked questions

Is Google Ads Brand Lift free?

The measurement itself is free, but the campaigns in the study must meet a minimum budget within 10 days, from $5,000 for one question in Country A markets to $60,000 for three questions in the United States and other Country B or C markets.

Why can't I see Brand Lift in my Google Ads account?

Brand Lift is not available to every account. Google says you need to contact your Google account representative to use it, and without one you cannot use the feature.

Which campaigns can use Brand Lift?

Video, audio and Demand Gen campaigns. For Demand Gen, only YouTube video ads are measured, not Gmail or Discover.

How long does a Brand Lift study take?

The budget must be met in the first 10 days, and studies usually finish within 14 days. Standard studies take 3 to 14 days and Enhanced Lift studies 9 to 14 days. A study can end early once it has about 4,100 responses per metric.

What is the difference between absolute, relative and headroom lift?

Absolute lift is the exposed positive rate minus the baseline rate. Relative lift divides that difference by the baseline rate. Headroom lift divides it by 1 minus the baseline rate. With rates of 20% and 40%, they are 20%, 100% and 25%.

For the formats and specs behind these campaigns, see YouTube ad specs, YouTube Shorts ads and Demand Gen campaigns.

Want a fast landing page for the people your video ads reach? See our website development service and all our services, or email [email protected].