Brand Lift: Studies, Metrics, Costs
Brand lift explained for buyers: how studies work, which metrics matter, Google and LinkedIn minimum budgets, sample sizes and setup errors to avoid.
Published 2026-11-01 · Video Marketing · By Tommaso Maria Ricci, Founder of Neverframe
A brand lift study on YouTube in the United States asks for a $10,000 budget inside ten days to measure one survey question, and Google can still hand back a report that says "No lift detected". Buyers often find this out after the media plan is signed. Brand lift is the measured change in what people think about a brand because they saw its advertising. Two matched groups get the same survey question: people who were served the ad and people who were deliberately held back from it. The gap between their answers is the lift. The test itself is free on Google, Meta and LinkedIn, so the real cost is the media budget needed to qualify and a creative that can move a number. Below are the metrics, the current minimums by platform, the sample sizes behind a readable result, and what to brief before the shoot.
What brand lift measures
PPC Land, the trade publication, puts the reason for the metric plainly: "Clicks, viewable impressions and last-click conversions record what happened on a screen, not whether a viewer now remembers the advertiser or would put it on a shortlist." A brand film can run for a month, collect millions of views and leave a marketing director with no answer to the question the CFO will ask in the budget meeting. A lift study is built to report whether anyone's opinion moved because of the ad.
Five questions come up in nearly every study, and they map to the funnel. Ad recall asks whether people remember seeing an ad from the brand recently. Awareness asks whether they have heard of the brand at all. Consideration asks whether they would put it on a list. Favorability asks how they feel about it, and purchase intent asks whether they plan to buy. Google lets you pick "up to 3" of these in its setup flow, along with association and a custom metric, and each extra question raises the budget you need, so the choice is a commercial decision as well as a research one.
Google's own definition is worth reading in full because it names the scope of the tool. In its help documentation, Google writes:
"Brand Lift is a free tool for measuring the effectiveness of your video ads, which you can use to adjust and improve your Video and Demand Gen campaigns."
The Google Ads Help page on Brand Lift explains that "Once your ad campaigns start running, Google will start showing your Brand Lift surveys on YouTube, before a video starts." One group saw your ads. The other group was eligible and did not get them. Anyone reading a lift report should hold on to that design, because everything that can go wrong later comes back to the purity of those two groups.
If you already track video analytics and KPIs, treat lift as a separate layer on top of them. A metric like view through rate describes attention to the ad, and a lift number describes what happened in the viewer's head afterwards. A video can hold attention and still leave no trace of the brand name.
Brand lift metrics compared
Most confusion in a lift readout comes from three percentages that describe the same result. Google uses one example across all of them: a positive response rate of 20% in the baseline group and 40% in the exposed group. The same two numbers produce an absolute lift of 20%, a relative lift of 100% and a headroom lift of 25%. Check which of the three a report leads with, since the 100% looks far bigger than the 20% and describes the same result.
| Metric | How it is calculated | Result for 20% vs 40% | Use it for |
|---|---|---|---|
| Absolute brand lift | Exposed rate minus baseline rate | 20% | Sizing the real change in percentage points |
| Relative brand lift | Difference divided by baseline rate | 100% | Comparing against a low starting point, with care |
| Headroom brand lift | Absolute lift divided by 1 minus baseline rate | 25% | Well-known brands with little room left to grow |
| Lifted users | Survey gap extended to campaign reach | Depends on reach | Scale of the effect |
| Cost per lifted user | Campaign cost divided by lifted users | Depends on spend | Comparing campaigns and creatives |
Google's recommendation on which of these should run the decision is unusually direct. Its guide to lift statuses and metrics says:
"Absolute lift doesn't necessarily reflect your overall brand lift performance. It is better to focus on a metric like cost per lifted user as the primary success metric of your campaign, because it factors in both reach and cost."
The same page runs a small example with two $100 campaigns. Campaign 1 buys at a $15 CPM, reaches 6,666 people and earns a 10% absolute lift, which works out to 667 lifted users at $0.15 each. Campaign 2 buys at a $5 CPM, reaches 20,000 people, earns half the lift and ends up with 1,000 lifted users at $0.10 each. Google draws the conclusion itself: "If you look at absolute lift only, Campaign 1 seems to perform better than Campaign 2. But at the same cost, Campaign 2 drove 50% more lifted users, at a 66% lower CPM, and with a 33% more efficient cost per lifted user."
There are two caveats on the lifted users number. Google notes that it "doesn't necessarily measure unique users", since one person can be counted more than once across a campaign. Reported lift is also an estimate from a sample, and Google is open about the range: "Usually, the confidence interval is 90%, so you can expect that in 90% of the cases, the true lift number will be within that range". When a vendor quotes a single figure with no range, ask for the interval before it goes into a board deck.
The cost per lifted user is also how a video team should compare creative. Two cuts of the same film can run on similar media and return very different costs per lifted user, and that difference is the most useful finding a study can produce. It belongs with the rest of your video marketing ROI reporting, since finance will ask how brand spending compares with the performance line.
Brand lift study costs by platform
No platform charges for the survey itself, which is where the phrase "free tool" misleads people. What you pay for is the qualifying media. The minimums are public on Google and LinkedIn. Meta and TikTok route most of the detail through an account representative, so any figure you read for those two on a third-party blog deserves suspicion. We found vendor pages quoting Meta and TikTok minimums that disagree with each other by a wide margin, and we left those numbers out of this guide.
| Platform | Cost of the test | Published minimum | Who can set it up | Duration |
|---|---|---|---|---|
| Google / YouTube | Free | US: $10,000 for 1 question, $20,000 for 2, $60,000 for 3 | Accounts with a Google representative | 3 to 14 days (Standard) |
| Google Enhanced Lift | Free | 3x the Standard minimum | Same | 9 to 14 days |
| Free | $60,000 for 1 question, rising with more questions | Campaign Manager | 2 weeks to 90 days | |
| Meta | Free to create | Not published, set by audience location | Account representative or brand survey test | Set per test |
| TikTok | Not published | Not published | Account manager only | Set per test |
Google's budget table depends on where you advertise. The Set up Brand Lift page sorts countries into three groups. The United States, the United Kingdom, France and Spain sit in the middle group. Australia, Canada, Germany, Italy and Japan sit in the highest one, at $15,000 for one question and $60,000 for three. A campaign across several countries takes the highest bucket among them, and the spend has to land within ten days of the study start. Access is its own hurdle, since "Brand Lift isn't available for all Google Ads accounts. You'll need to contact your Google account representative to use Brand Lift."
The cheapest way to pass Google's threshold is to measure less. In the setup guide, Google says "The budget minimum goes up or down based on the number of questions you have. By removing questions, you lower the minimum budget required." Going from two questions to three in the US triples the requirement from $20,000 to $60,000, so a third question has to justify $40,000 of committed media on its own. Most mid-size brands should choose one metric and defend the choice.
LinkedIn sets the bar higher and states it clearly. Its Brand Lift Testing guide says "While there's no cost for Brand Lift Testing, you need to meet a minimum cumulative budget across your ad account's campaigns to set up tests." The next sentence gives the number: "For example, for one question, you need a minimum of $60,000 USD or equivalent currency to qualify, and that minimum changes as you add additional questions." Audience size matters as well: "While conclusive results aren't guaranteed, we recommend an audience size of at least 1 million and strongly encourage a minimum audience size of at least 300,000". A B2B brand targeting a narrow job title may not be able to meet that, whatever it spends.
On Meta, the About brand lift tests page describes the entry point this way: "The test is free to create, but the advertising campaigns and ad sets that you use for your test may have certain requirements to ensure quality results." TikTok keeps it shorter, and the TikTok Brand Lift Studies page is explicit: "TikTok Brand Lift studies can only be set up by your account manager." Below a certain spend, independent research firms are the remaining route. Nielsen Brand Lift is one example, and the company describes its method as "survey-based control/exposed techniques". Its fees sit outside any platform's free offer and need a quote.
How many survey responses a result needs
A strong film that moves ad recall by five points shows up quickly. A campaign that moves consideration by half a point can burn through a six-figure budget before the statistics catch up. Google publishes the ranges, and its Display & Video 360 documentation carries nearly the same table, with a small difference at the 2% row.
| Detectable absolute lift | Total survey responses required (Google Ads) |
|---|---|
| Above 4% | 1,200 to 2,800 |
| 3% | 2,800 to 5,000 |
| 2% | 4,100 to 11,000 |
| 1.5% | 11,000 to 20,000 |
| 1% | 20,000 to 45,000 |
| 0.5% | 45,000 to 180,000 |
| Below 0.5% | More than 180,000 |
Those ranges come with a warning that buyers should read before committing budget:
"For high-performing campaigns, you can expect to detect lift once you receive about 2,000 responses per lift metric. At the recommended budget minimum, you can expect to detect lift once you receive 4,100 responses per lift metric. If your campaign has not shown any lift after reaching 16,800 responses per metric, you may not be able to detect your lift."
Google's status column follows fixed progress bands. Under 50% of study progress, Google shows "Not enough data". Between 50% and 100%, it reports statistically significant positive lift if it finds any. At 100%, a result that never crossed the line turns into "No lift detected". Google also says "It's recommended waiting until the study is complete for the final report, because the results may change throughout the duration of the study." LinkedIn gives similar advice about the period after the flight: "Result metrics can fluctuate for up to 15 days after the end date of the test." Its methodology also changed recently. The statistical strength page says "As of June 2026, we're changing our measurement framework from Frequentist to Bayesian methodology", so results from before and after that date are presented differently and should not be stacked in one trend chart.
If you already have a study that keeps returning nothing, Google offers Enhanced Lift, which needs three times the Standard minimum and collects roughly three times the surveys. It raises the odds without promising anything, and the setup guide says as much: "While Enhanced Lift greatly improves the chances of detecting lift, it doesn't guarantee positive results." Before tripling the budget, review the creative, because extra responses measure the existing film and leave it unchanged.
Setup errors that invalidate a brand lift study
Lift research has a long record of failing for mundane reasons. The IAB said so in its 2012 checklist for ad effectiveness studies, written when most studies were still run by outside vendors on banner placements: "Ad effectiveness studies are complicated, there are a lot of moving parts and simple errors can invalidate the results of a study". The media has changed since then and the logic has held. The checklist's standard for the control group still applies: "True experimental design (random recruitment of control/exposed from the same placements) is the gold standard." It also states the principle every platform now builds on: "The only differences between the control and exposed groups should be exposure to the campaign".
Contaminated holdouts are the most common way to break that principle. Meta's warning is specific:
"You shouldn't use Brand Lift if you have other active campaigns for the same brand or product running across Meta technologies. Running more than one campaign may result in people in the holdout group seeing your ad from one of the other campaigns, which would contaminate the results."
TikTok asks for the same discipline in its considerations list: "Include all active TikTok campaigns in your study." Google's version of the problem shows up with retargeting and sequences. Its documentation says "If a Brand Lift study uses campaigns that target audiences that viewed the ad video before, it currently does not build a control group." A study on a retargeting line can make progress on the exposed side, and then never report anything.
Run this list with your media team before the flight starts.
| Check | What to confirm | What breaks if you skip it |
|---|---|---|
| One question, one job | The metric matches the campaign's goal | You buy a result nobody will act on |
| Spend, not just budget | Actual delivery meets the minimum inside the window | Status stays on "Not enough data" |
| Other campaigns | No parallel activity for the same brand on the platform | Holdout sees the ad and the gap shrinks |
| Targeting width | Country-level or broad audience where possible | Too few viewers to survey |
| Retargeting | No lists of past viewers inside the study | No control group gets built |
| Survey language | Same language as the ad and the targeting | Viewers dismiss the survey |
| Brand name in answers | Specific product name that matches the ad | Respondents cannot pick the right answer |
| Competitor list | Real alternatives viewers would recognize | The answer set looks fake and skews responses |
| Creative lock | Same cut and budget through the flight | You cannot tell what caused the change |
| Read date | Decision meeting set after the study closes | Early numbers get quoted and later reversed |
Narrow targeting deserves its own note, because brand teams love it. Google warns that "More restrictive targeting types, such as placements, keywords, and retargeting, reduce the number of eligible viewers and can lead to less impressions." It adds that studies are ideally run at the country level. The language rule catches international teams in the same way: "The survey question language should match the language in the video or audio ad associated with your campaign."
Brand lift vs conversion lift
These two tests get mixed up in briefs because they share a design. Both hold out a control group, and they answer different questions. LinkedIn's documentation describes the pair side by side, saying that "Conversion Lift Testing - Helps you determine how effective your campaigns are at driving conversions." Brand lift reads a survey. Conversion lift reads actions such as purchases and sign-ups recorded by the platform.
| Brand lift | Conversion lift | |
|---|---|---|
| Data source | Survey answers from exposed and control groups | Recorded conversions from test and control groups |
| Question answered | Did opinion or memory change? | Did the ads cause extra actions? |
| Best for | Brand films, launches, awareness flights | Performance creative and direct response |
| Time to effect | Days after exposure | Depends on the sales cycle |
| Main weakness | Stated answers can differ from behavior | Long sales cycles escape the window |
Choosing between them is mostly a question of what the campaign was bought to do. A launch film for a B2B brand with a nine-month sales cycle will not show its value in a conversion window, so a lift study is the honest instrument. A retailer running bumper ads against a seasonal offer usually wants conversion lift, with brand lift added when the budget clears the minimum. LinkedIn also blocks the two from running together at account level, so plan the order if you want both: "an account-level Conversion Lift test and an account-level Brand Lift test can't run simultaneously in the same account."
Why the creative decides the result
Kantar's reading of its own database, published in its analysis of French Cannes winners, is that "Creative quality contributes to nearly 50% of media impact far surpassing reach or frequency of touch-points, according to Kantar CrossMedia database." A separate piece of research points the same way on a different outcome. An Adweek sponsored article reports that "Recent research from NCSolutions and Nielsen said creativity provides 47% of the total sales impact." The first figure concerns media impact and the second concerns sales, so they back up each other's direction without measuring the same thing.
Platforms say this too, in their own dry language. LinkedIn's best practices for Brand Lift tests include "Ensuring your ads have engaging content" on its list of ways to strengthen a result, and TikTok names comparing creatives among the benefits of its study. For that reason we would plan a lift study from the first script meeting, with the survey question on the table next to the brief.
A few creative decisions tend to show up directly in the readout. These are working rules from production, and you should check them against your own results.
- Name the brand early and more than once. Ad recall questions ask about a specific brand. A beautiful film that reveals the logo in the last two seconds gives skippers nothing to remember, and a good hook rate will not save it if the hook never carried the name.
- Build for the slot the survey follows. On YouTube the placement is often skippable, so the first five seconds carry most of the load. Our pre-roll ads guide goes through the formats and lengths.
- Match the message to the question. If the study asks about consideration, the film has to give a reason to consider. A mood piece can lift favorability and leave consideration flat.
- Plan more than one cut. Google lets you split results by campaign, and cost per lifted user can differ sharply between a 30-second spot and a 15-second ad. Without cuts to compare, you learn whether the campaign worked and nothing about why.
- Watch frequency during the flight. Repetition helps recall up to a point. Our guide to ad fatigue covers how to spot when it starts working against you.
Testing before the flight lowers the risk of a very expensive "No lift detected". A cheap video A/B test on hook and brand timing can pick the stronger cut for far less than a lift study's minimum.
How a studio like Neverframe fits into a lift plan
Neverframe is a production studio, so we make the films and leave the survey to the platform or your research partner. We work in four formats for brands and agencies, and each one maps to a different lift job.
| Format | What you get | Timing | Typical lift goal |
|---|---|---|---|
| Films | A 60 to 90 second brand film with original script, cinematic score, 4K master and social cuts | Three to four weeks | Favorability, association |
| Spots | A 30 second spot with script, storyboard, licensed music, sound design and voice over, 4K master plus vertical version | Two weeks | Ad recall, awareness |
| Social ads | Three 15 second ads with different hooks plus two 6 second bumpers, in 9:16, 1:1 and 16:9 | Two weeks | Ad recall, comparing hooks |
| Avatars | A digital double of your CEO who records new messages without new shoots | First version in days | Message association in updates and outreach |
First versions arrive in a few days, which gives you time to run a small test on the cuts before the budget for a study is committed. Agencies get a first version in 72 hours and can work with us under white label through the partner program. If a lift study is on your calendar this quarter, see the four formats and pick the one that matches the metric you plan to measure.
A simple plan for your first brand lift study
Start at least six weeks before the flight if the film does not exist yet. Pick one metric in a sentence that the CFO would accept, such as we want more US marketing directors to recall this brand after seeing the film. That sentence fixes the question, the audience and the country bucket, and those three give you the minimum budget. Check the budget against the public table for Google, or ask your representative for Meta and TikTok, and confirm access before anyone writes a script.
Next, make the creative with the survey in mind, and cut at least two versions. Google's review step for surveys can take up to 48 hours, so build that into the launch. During the flight, leave the creative and budget alone and watch actual spend against the minimum, since a budget that does not deliver counts for nothing. After the study closes, read the absolute lift with its confidence range, then compare cost per lifted user across cuts. Keep the file, because a benchmark from your own brand and category is worth more than any industry average, and your second study will be cheaper to plan.
Ready to brief the film that goes into the study? Tell us about the campaign with the metric, the market and the launch date, and we will come back with the format and schedule that fit.
FAQ
What is brand lift in simple terms?
Brand lift is the difference in opinion or memory between people who saw an ad and similar people who did not. A platform splits an eligible audience at random, holds one group back, and asks both groups the same survey question. If 40% of the exposed group answers positively against 20% of the control group, the absolute lift is 20%. Because the two groups differ only in exposure to the ad, the gap can be credited to the campaign, which is something views and clicks cannot tell you.
How much does a brand lift study cost?
The survey is free on Google, Meta and LinkedIn, and the cost sits in the qualifying media. Google asks for $10,000 over ten days for one question in the United States, $20,000 for two and $60,000 for three. LinkedIn asks for a minimum of $60,000 for one question. Meta sets requirements by audience location through its representatives, and TikTok studies go through an account manager. Independent research firms such as Nielsen run their own studies at quoted fees, so ask for a price before you plan around one.
How long does a brand lift study take?
On Google, a Standard study completes within 3 to 14 days depending on how fast surveys are collected, and Enhanced Lift takes 9 to 14 days. Google allows up to ten days to meet the spend requirement and may keep collecting responses for a short period after that. LinkedIn tests run between two weeks and 90 days, and LinkedIn recommends waiting 15 days after the end date before treating results as final. Allow time for survey approval too, since Google's review can take up to 48 hours.
What is a good brand lift result?
There is no universal benchmark worth quoting, since results depend on category, brand size, metric and starting awareness. Treat any single industry average with suspicion. Use your own history where you have it, and compare campaigns on cost per lifted user, which Google recommends as the primary success metric because it accounts for both reach and cost. As a practical signal on sample size, a 2% absolute lift needs roughly 4,100 to 11,000 responses on Google to be detected, while a lift above 4% needs far fewer.
Why did my brand lift study show no lift?
A "No lift detected" status at 100% progress means the study collected its responses and found no statistically significant positive difference. The usual causes are a creative that never makes the brand memorable, a control group contaminated by other campaigns, or a metric that does not match the message. "Not enough data" is a separate status and usually points to underspending, narrow targeting or a low bid. Check delivery and setup first, then look hard at the film before paying for a bigger study.