The short answer
A platform brand lift study is run by the platform that sold the media, on its own inventory, with a holdout the platform controls. An independent study is run by someone who doesn’t sell the media, can cover every channel including offline, and shows its working. Platform studies can be fine for one digital channel; independent studies are needed to compare channels or prove a mixed campaign worked.
Key takeaways
- Platform studies only see their own inventory, and the platform controls the holdout and the reporting.
- Google says lift is typically detectable at around 2,000 responses per metric for strong campaigns.
- Independent studies can measure offline and mixed media and compare channels on one method.
- Use platform tools for in-channel optimisation; use independent studies to prove what the whole campaign did.
How platform brand lift works.
Large ad platforms offer brand lift studies inside their own tools. People eligible to see your ads are split into an exposed group and a holdout, and short surveys are shown to both. Google’s Brand Lift reports ad recall, awareness and consideration, with lifted users and cost per lifted user.
These studies need scale. Google’s guidance says high-performing campaigns can expect to detect lift at about 2,000 responses per metric, around 4,100 at the recommended budget minimum, and that detecting a 2% absolute lift needs 4,100 to 11,000 responses.
The independence problem.
The platform that sold the media chooses the holdout, runs the survey on its own inventory and reports the result. Advertisers have long described this as platforms grading their own homework. It also means each platform measures only its own slice: a platform study can’t tell you what your TV, outdoor, print or retail media did, or compare channels on the same method.
In TransUnion and EMARKETER’s 2025 survey, 41% of US marketers named walled-garden reporting limits as a top barrier to accurate measurement.
Side by side.
| Platform brand lift | Independent brand lift | |
|---|---|---|
| Who runs it | The platform that sold the media | A third party that doesn’t sell media |
| Channels | That platform’s inventory only | Any channel, including offline and mixed |
| Holdout | Set by the platform | Matched control, designed and documented |
| Data access | Aggregated metrics | Raw data, exclusions and method (with SpinProof, the Evidence File) |
| Best for | Optimising inside one digital channel | Proving what a campaign did and comparing channels |
When to use which.
- Use a platform study to optimise creative or audiences inside a single digital channel where you already spend at scale.
- Use an independent study when the campaign runs across channels, includes offline media, needs to be compared with other channels, or has to stand up to a finance director or an agency review.
- Use both when you want in-channel optimisation and an independent answer to whether the whole campaign worked.
Questions
Are platform brand lift studies free?
They are usually free to run but need media spend above a minimum, and the platform sets the method. Independent studies cost money but work across channels and show their working.
Can an independent study measure a digital campaign too?
Yes. Independent studies cover digital channels as well as offline ones, which is what lets them compare channels on one method.