← All Labs

Brand Lift

Flagship
Measurement
01

Customer question

Did advertising change how people think about our brand?

02

What this solution does

You ask people who saw the ads, and people who didn't, the same questions about the brand — the gap between their answers is the campaign's effect on perception.

Specifically, it answers: Did a campaign move perception metrics like awareness, consideration, or purchase intent, and by how much, compared to people who weren't exposed to it?

03

How it works

An exposed group (saw the ads) and a control group (didn't) are surveyed on the same brand perception questions; the difference in responses between groups is the estimated lift attributable to the campaign.

04

Data requirements

  • A defined exposed population (saw the campaign)
  • A comparable control population (did not see the campaign)
  • Survey responses on brand perception metrics from both groups
  • Sufficient sample size in both groups for statistical reliability
05

Discovery questions

How will exposed and control groups be defined and separated?

Poor separation between groups (e.g. control group partially exposed) biases the lift estimate toward zero.

What sample size is achievable in each group?

Small samples produce wide confidence intervals, making it hard to distinguish real lift from noise.

Is this measured via an independent research panel or a platform-native tool?

Platform-native brand lift tools use the platform's own exposed/control logic, which can differ from independent methodology and carries different bias considerations.

06

Build an experiment

07

Generate synthetic data

08

Data readiness

Generate a dataset in the previous step to run readiness checks.
09

Run model / simulation

Generate a dataset first to run the model.
10

Results

Run the model in the previous step to see results here.
11–12

Business ⟷ Technical interpretation

Awareness moved from controlAwareness% (control) to exposedAwareness% (exposed) — a awarenessLift percentage point lift — with purchase intent showing a smaller but still positive intentLiftpp lift.

Shown as a template — the highlighted tokens are filled in once the model has been run.

13

What would you do next?

  • Extend the study if confidence intervals are too wide to act on confidently
  • Compare independent brand lift results against any platform-native brand lift reporting for the same campaign
14

Solution & vendor landscape

Implementation approaches

Independent brand lift research vendorPlatform-native brand lift studies (e.g. within a media platform)Media agency-run brand tracking
See the full Vendor Landscape →