Brand Lift Study: The Straight Explanation

What a Brand Lift Study Actually Measures

A brand lift study measures whether people who were shown your ad changed how they think or feel about your brand, compared to people who weren't shown it. It's a survey, not a sales report. Google runs it on YouTube and Demand Gen campaigns; Meta runs its own version across Facebook and Instagram. Both ask the same basic question in different clothes: did the ad move perception, and by how much?

The mechanics are the same everywhere. The platform splits your audience into two groups: an exposed group who saw the ad, and a control group who were eligible to see it but didn't, usually because of how the auction happened to serve. Both groups get shown a short survey on whatever metric you picked when you set up the study — awareness, ad recall, purchase consideration. The gap between how the two groups answer is the "lift."

That's the whole idea. It's a controlled comparison of stated opinion, run at platform scale, at no separate tooling cost on either Google or Meta. It tells you about brand perception. It does not, on its own, tell you whether anyone actually bought anything.

How the Numbers Actually Work

Say a study surveys 10,000 people who saw the ad and 10,000 who didn't. In the control group, 8% say they'd consider the brand — that's the baseline positive response rate. In the exposed group, 11% say the same — the exposed positive response rate. The difference, 3 percentage points, is the absolute brand lift. Divide that by the baseline (3 divided by 8) and you get roughly 37.5% relative brand lift — the number most dashboards lead with because it looks bigger.

From there the platform estimates lifted users — how many people, scaled across your actual campaign reach, moved from "no" to "yes" because of exposure — and cost per lifted user, media spend divided by that number. If the campaign cost $40,000 and the model estimates 2,500 lifted users, that's $16 per lifted user. None of these numbers touch a shopping cart. They're all downstream of a survey answer, and that's fine as long as you remember what question they're answering.

Brand Lift Study on YouTube

Google's version runs on Video and Demand Gen campaigns and needs a few things most people don't expect going in. It isn't self-serve: you need a Google account representative to turn it on — there's no toggle for it in a standard self-managed account. Once it's running, there's a minimum spend requirement over a 10-day measurement window, and the eligibility calculator can flip a study to "Not Eligible" if a campaign pauses, ends, or falls under pace, sometimes with a lag of a few hours before the status catches up with reality.

The common mistake is pulling budget the moment progress reads close to 100% and looks "on track." Don't cut budget until you've actually crossed the line — pacing dips can revert a study to ineligible and cost you the window. If a study goes sideways — ineligible before it started, or a partial run with almost no survey responses — the fix isn't to nurse it along. Stop it and start a clean one with the full 10 days ahead of it.

Brand Lift Study on Meta

Meta's brand lift studies work on the same exposed-versus-control logic but sit inside Ads Manager rather than a YouTube-specific tool, and larger budgets are typically set up with a Meta rep rather than opened self-serve. The survey wording and question set differ from Google's, and the audience split happens across Facebook and Instagram placements rather than pre-roll video slots. The underlying math — baseline response, exposed response, absolute and relative lift — is the same idea wearing a different platform's interface.

Where people get tripped up on Meta is treating a positive brand lift result as proof the campaign paid for itself. It answers a narrower question: did the people who saw this specific creative say something different about the brand than the people who didn't. That's useful. It's not a revenue number, and Meta doesn't present it as one.

What a Brand Lift Study Can't Tell You

Here's the gap most explainers on this topic skip: a brand lift study measures stated perception, sampled from survey respondents, not purchase behavior. It can show that more people say they'd consider your brand. It cannot show that display or video drove an incremental sale, because nobody in the control group is tracked through to a purchase — they're tracked through to a survey click.

We've made this mistake ourselves: reporting a strong relative lift number to a client as if it settled the incrementality question, when what it actually settled was a perception question. The two get conflated constantly because both use the word "lift" and both compare an exposed group to a control group. The methodology that answers the sales question is a holdout test or matched-control study, measured against actual attributed purchases rather than survey responses. It's a different tool solving a different problem, and most advertisers need both, not one instead of the other.

Where This Leaves Amazon Advertisers

Amazon doesn't publish a self-serve "brand lift study" tool the way Google and Meta do. There's no survey product sitting inside DSP the same way. The nearest equivalent — holdout groups and matched controls reconciled in Amazon Marketing Cloud — measures something closer to the sales question than the perception one: whether display and video actually drove purchases that wouldn't have happened otherwise, without DSP and sponsored ads double-counting the same shopper.

That's the product Dr. DSP is built around — Dr. DSP is Amazon's DSP, the demand-side platform for programmatic display, video and audio, not the Delivery Service Partner courier franchise. It's run as a managed product by Fable 5, from Full Circle, which has managed more than $500M in revenue across 100+ brands. Across 30 of those advertiser accounts in July 2026, the book produced a 6.04x return on ad spend against 78.4 million impressions at a $4.00 CPM — numbers built on attributed purchases, not survey sampling.