Brand Lift and Search Lift on YouTube in 2026: Measuring Video Without Clicks
Your video campaign served 400,000 impressions, produced 11 clicks and two view-through conversions. Your finance lead asks what that bought, and the only answer available is “it builds awareness”. The problem is not that video fails to work. The problem is that you are measuring it with a tool designed for clicks.
Brand Lift and Search Lift are two studies inside Google Ads that answer “did the video change behaviour” without depending on a click. This article covers how the measurement works, the budget thresholds involved, how to read lifted users and cost per lifted user, how Search Lift differs from Brand Lift, where the methodology breaks, and what to run instead when you lack the volume for a formal study.
Why clicks are the wrong metric for video
Three reasons the click report undersells a video campaign:
- The format is not built for clicking. People watch YouTube on a TV or with the phone face down. There is often nothing to click and no reason to — the response happens later, somewhere else.
- View-through attribution is contested. A conversion credited after an unclicked view might have happened anyway. That is a hypothesis, not a measurement.
- The effect is delayed. Days or weeks pass between the view and the purchase, and the credit lands on the last channel — usually the branded search you yourself warmed up.
Video rarely closes the deal. It changes the probability that someone enters your funnel at all. Measuring that with clicks is like grading a billboard by how many people walked into it.
How Brand Lift works
Brand Lift is a controlled experiment with a survey attached:
- The audience splits into an exposed group that saw your ad and a control group that was eligible but was not shown it.
- Both groups receive a short single-question survey inside the YouTube interface.
- The difference in correct answers between the groups is the lift.
The critical detail: the control group is drawn at random from users who could have been served your ad. That randomisation is what makes the result causal rather than correlational — unlike any before-and-after comparison.
What you can measure
| Metric | The underlying question | When it fits |
|---|---|---|
| Ad recall | Do you remember seeing an ad for this brand? | Always: the baseline metric, and the fastest to move |
| Awareness | Do you know this brand? | New brands, entry into a new market |
| Consideration | Which brand would you consider buying? | Competitive categories where people compare |
| Favorability | Which brand do you prefer? | Repositioning, reputation work |
| Purchase intent | Which are you likely to buy? | Short cycles, mass-market products |
Rule of thumb: the further down the funnel a metric sits, the harder video moves it and the more budget it takes. Ad recall almost always lifts; purchase intent frequently does not, and that is a legitimate result rather than a failed campaign.
Reading the numbers
- Absolute lift is the difference in percentage points. Exposed 26%, control 20% means a 6-point absolute lift.
- Relative lift expresses the same gap against the baseline: 6/20 = 30%. It looks better in a deck and misleads more easily when the baseline is small.
- Lifted users estimates how many people changed their view because of the ad.
- Cost per lifted user is spend divided by lifted users — the only figure that lets you compare creatives and campaigns fairly.
- Headroom tells you how much room is left: if 78% of the control group already knows the brand, further awareness gains will be expensive.
When comparing creatives, judge on cost per lifted user. A video with a smaller absolute lift but a third of the cost per lifted user is the better candidate for scale.
Search Lift: behaviour instead of opinion
Search Lift measures whether brand- or product-related searches increased among the exposed group relative to control. It answers “did people go looking for us after watching?”
| Brand Lift | Search Lift | |
|---|---|---|
| Data source | User survey responses | Actual search activity |
| What it shows | Change in perception | Change in behaviour |
| Speed to result | Usually faster | Needs search sessions to accumulate |
| Weak spot | People answer surveys imprecisely | Seasonal demand can muddy the read without a clean control |
| What you do with it | Rework the message | Prepare branded search for the incoming demand |
Search Lift earns its place because it links video to what happens next: rising branded queries mean you need to reinforce the brand campaign before the demand you paid for is harvested by a competitor. The economics of that defence are covered in our article on impression share and auction insights.
Thresholds: when a study will actually run
Google adjusts the exact minimums and they vary by market, so confirm current requirements in the interface. Useful planning benchmarks:
- Budget. A study needs meaningful spend inside a compressed window — typically thousands of dollars across 7–10 days, not a trickle of video in the background.
- Timeframe. Slow, drawn-out delivery dilutes the sample. A concentrated flight works better.
- One variable. If creative, geo and targeting all change mid-flight, the result explains nothing.
- Statistical significance. A result that never reaches significance is not “a small lift” — it is an absence of an answer.
If the budget is not there, do not try to squeeze an answer out of a micro-study. Run a geo experiment instead: enable video in one set of regions, withhold it in a comparable set, then compare total conversions and branded query volume. The methodology is detailed in our guide to experiments in Google Ads.
Running a study step by step
- Write the hypothesis before launch. “A product demo will lift consideration more than the brand film” is a hypothesis. “Let’s see what happens” is not.
- Pick one primary metric. Awareness or ad recall for a new brand; consideration in a competitive category.
- Keep the survey question neutral. Leading phrasing distorts both groups, the exposed one most of all.
- Freeze the campaign during the study. Any targeting or budget edit breaks comparability between the groups.
- Plan for steady delivery. Choppy pacing produces a muddy sample.
- Prepare the lower funnel first. Brand campaign, remarketing to viewers, landing page that matches the promise in the video.
- Read lift alongside business metrics. Lift without sales points to the offer or the funnel, not to the video.
Where the methodology breaks
- Baseline too high. If 80% already know the brand, there is little awareness left to lift. Choose a metric further down the funnel.
- Overlapping campaigns. Simultaneous TV, out-of-home and influencer activity makes lift attribution arguable.
- Targeting too broad. The survey reaches people outside your category and the lift dissolves into the sample.
- No brand in the opening seconds. Ad recall measures brand recognition; if the logo appears at second 25 and average view time is eight seconds, there is nothing to measure.
- Chasing relative lift. A 200% relative lift on a 1% baseline is statistical noise in a nice suit.
Three results and what to do with them
| Result | Diagnosis | Action |
|---|---|---|
| Ad recall up, consideration flat | The ad was noticed but the argument did not land | Rewrite the message: less mood, more concrete reason to choose you |
| Lift achieved, branded search unchanged | No bridge to the next step | Add a clear call to action and a searchable brand cue; switch on viewer remarketing |
| No lift at all | Creative, audience, or a baseline with no headroom | Check the first five seconds, tighten the audience, change the metric |
In practice, the audience for video and its downstream remarketing is easiest to build with custom intent segments, while reach expansion should be controlled the way we describe in our guide to optimized targeting and audience signals. If lift stalls because of the creative, start with the asset set — see Ad Strength and asset performance reporting.
Folding video into the account’s reporting
The mistake is quarantining video in a separate world of “brand metrics”. A workable structure:
- Video is accountable for lift (Brand Lift / Search Lift) and for growth of the remarketing pool.
- Branded search and remarketing are accountable for converting the demand video created.
- The overall check is incrementality: toggle video by region and read the account’s total result, not one campaign’s.
- Economics are calculated at channel level, never on last click.
How the video-to-demand-to-conversion chain plays out in practice is covered in our article on video ads for performance on YouTube and Shorts, and the demand-creation side in our guide to Demand Gen in Google Ads.
If the goal is a repeatable media-buying process rather than a one-off study, PPC Rebels offers agency ad accounts for high-volume advertisers, and the English blog archive covers the measurement stack around it.
Creative built for lift: what actually moves the number
A study measures the video, not the media plan. When there is no lift, nine times out of ten the culprit is the opening seconds rather than the targeting. What to check, in order:
| Element | Practical rule | What breaks when ignored |
|---|---|---|
| Brand on screen | Logo or name within the first 3–5 seconds | Ad recall: people remember an ad but not whose it was |
| One idea per video | A single argument, not five benefits in a row | Consideration: nothing sticks |
| Sound as an option | The point lands without audio; key lines are on screen | A large share of views happen muted |
| Problem before solution | The first five seconds show a situation the viewer recognises | The skip happens before the argument arrives |
| Format per surface | Vertical for Shorts, landscape for connected TV | Cropped frames and lost on-screen text |
On length: short formats usually lift recall better because they get completed; longer ones move consideration because they have room to make a case. The sensible approach is one short cut for the top of the funnel and one longer cut for consideration, rather than a single universal edit that does neither job well.
Preparing a creative set for a study
- Take one hypothesis and produce two or three variants that differ by one element: the hook, the length, or whether an offer closes the spot.
- Do not pair different variants with different audiences, or you end up measuring targeting instead of creative.
- Compare variants on cost per lifted user, not on view-through rate — completion and lift correlate weakly.
- Scale the winner and dissect the loser: was it the hook, the argument, or a missing brand cue?
Video economics without last click
The finance question is always “what did we earn”. For video the answer has two stages, and you need to state them before launch — otherwise every conversation ends in a comparison with branded search that video loses by definition.
A four-step model:
- Measure the demand created. Search Lift, or growth in branded query volume during the flight versus a comparable period without it.
- Measure what that demand converts into. Branded search and viewer remarketing are ordinary, fully measurable campaigns.
- Measure incrementality. A geo experiment: regions with video against comparable regions without, comparing total account conversions.
- Roll up channel economics. Video spend plus the harvesting spend against the revenue the increase produced.
An illustrative structure for that calculation: $12,000 spent on video; branded queries up by 3,400 for the month; 2,100 of those clicks reached the brand campaign; a 6% conversion rate produced 126 orders; at a $340 average order value and 30% margin that is $12,852 in contribution. Read: the channel roughly broke even on first purchase, and profit begins with the second — which makes the decision a question of LTV, not of a single transaction. How to calculate that boundary is covered in our guide to unit economics and allowable CPA.
Video is not supposed to pay for itself inside the video campaign report. It is supposed to pay for itself in the account report — and that is a conversation to have with the client before launch, not after.
A 30-day plan from launch to verdict
| Window | What you do | What you check |
|---|---|---|
| Week −1 | Prepare the lower funnel: brand campaign, viewer audience, landing page matching the video’s promise | Branded queries are not being handed to competitors |
| Days 1–3 | Launch, steady delivery, no edits | Even pacing, no sudden delivery gaps |
| Days 4–10 | Survey responses accumulate | Stay out of it; watch spend pacing only |
| Days 11–14 | First read of the result | Was significance reached, and which metric moved |
| Days 15–21 | Watch branded search and remarketing | Did demand travel down the funnel |
| Days 22–30 | Roll up economics and decide on scale | Cost per lifted user for each creative |
One more piece of advice: do not run a study during a major sale week. The demand spike lifts both groups and makes the read useless — you will be measuring the season, not the advertising.
FAQ
Does Brand Lift cost extra?
The study itself is not billed separately — you pay for media. But there is a budget and duration threshold below which the study will not run.
How long until results appear?
Early data usually shows within days of launch; the full read arrives once enough survey responses accumulate. Draw no conclusions before statistical significance is reached.
Can I measure several metrics at once?
Technically yes, but every extra metric splits the sample and raises the budget requirement. One primary metric per study is the practical choice.
How is Brand Lift different from Conversion Lift?
Brand Lift measures perception change through surveys; Conversion Lift measures the incremental conversions of the exposed group versus control. The second is closer to money but needs far more volume.
What if there is no lift?
Check three things in order: is the brand visible in the first five seconds, does the audience match the category, and is the baseline already too high for the metric you chose.
Does this work for small businesses?
A formal study rarely does, because of the thresholds. The substitute is a geo experiment: run video in one set of regions, withhold it in comparable ones, and compare total leads and branded queries.
Does video length affect lift?
It does, but not linearly. Short formats typically lift recall better; longer ones lift consideration because they have room to make the argument. Test rather than follow a rule.
Is lift measured across all devices?
Surveys appear wherever the user consumes YouTube, including mobile and connected TV. Connected TV delivers reach with no click possible at all, which is exactly why lift metrics matter there.
Can I compare lift across countries?
Carefully. Baseline awareness and survey-response culture differ by market. Compare trends within a market rather than absolute values between markets.
How do I connect lift to sales?
Not directly — they are different measurement layers. The chain runs through intermediate steps: lift, then branded search growth, then conversions from brand and remarketing campaigns.
Should I pause other campaigns during the study?
No need to pause, but large simultaneous launches in other channels complicate interpretation. Ideally choose a window without major offline activity overlapping.
What should I do immediately after a successful study?
Scale the creative with the lowest cost per lifted user, reinforce the brand campaign for the incoming demand, and switch on remarketing to viewers — otherwise the demand you created goes to whoever ranks first when those people search.
Bottom line: video does not owe you clicks — it owes you a change in behaviour. Brand Lift shows the change in perception, Search Lift shows the change in action, and a geo experiment remains the honest substitute when a formal study is out of budget reach.
