PSA control group testing
A PSA (public-service announcement) control test is an incrementality design where the control group is served unrelated placebo ads instead of the test campaign. Because both groups receive an ad impression, exposure conditions are similar, and the difference in conversions estimates the test campaign's incremental effect. It is an older alternative to ghost ads.
What this means
In a PSA control test, the experiment randomizes users into a treatment group that sees the real campaign and a control group that sees a placebo public-service announcement in the same slot. Serving an ad to both groups keeps the comparison closer to like-for-like than showing nothing to the control, since the mere fact of seeing an ad is held roughly constant.
The incremental effect is the difference in the outcome — conversions, visits, sign-ups — between the two exposed groups, attributable to the campaign content rather than to ad exposure in general.
Trade-offs versus ghost ads
The PSA approach has two costs: budget is spent serving placebo creative, and the PSA itself may have some effect on behavior, slightly biasing the baseline. Ghost ads were developed partly to address these issues by logging would-have-been-served control users instead of buying placebo impressions.
PSA tests remain a recognizable, intuitive incrementality design and are still used where a tangible control impression is desirable. As with all lift methods, they measure aggregate causal effect over a window, need adequate scale, and are distinct from credit-assigning attribution models.
- Control group sees a placebo PSA instead of the campaign
- Holds 'saw an ad' roughly constant across groups
- Costs placebo budget; ghost ads avoid that
How it appears in analytics and logs
A conversion-rate gap between users shown the real ad and users shown the PSA estimates incremental effect, controlling for the act of seeing an ad at all.
Diagnostic use case
Use a PSA control test to give the holdout group a comparable ad experience while measuring how many extra conversions the real campaign produced over the placebo baseline.
What WebmasterID can help detect
WebmasterID's first-party conversion events can be the outcome compared between the real-ad and PSA-exposed groups in such a study.
Common mistakes
- Assuming the PSA placebo has zero effect on behavior.
- Confusing the incremental result with attribution credit.
- Running the test below the scale needed to detect an effect.
Privacy and accuracy notes
PSA control tests compare aggregate group outcomes rather than profiling individuals. This page is educational and not statistical or legal advice.
Related pages
- Ghost ads methodology
Ghost ads are an experimental design for measuring ad effectiveness. Rather than showing a placebo ad to the control group, the system records which control users would have been served the test ad had they been in the treatment group, then compares only comparable users. This isolates the ad's incremental effect while avoiding the cost and bias of serving placebo creative.
- Conversion lift studies
A conversion lift study randomizes users into a group eligible to see ads and a control group held out from them, then compares conversion rates between the two. The difference estimates incremental conversions — those caused by the ads rather than ones that would have occurred anyway. Major ad platforms offer lift studies as a counterfactual check on attributed conversion counts.
- Incrementality testing: what attribution cannot tell you
Incrementality testing measures the lift a channel actually causes by withholding it from a control group and comparing outcomes. It answers the question every attribution model dodges: would this conversion have happened anyway? It is causal where attribution is merely correlational, but it requires deliberate experiment design.
- Attribution analytics
Compare placebo-controlled lift with attributed conversions.
Sources and verification notes
- Lewis & Rao — On the Near Impossibility of Measuring the Returns to AdvertisingDiscusses PSA/placebo control designs and their statistical challenges.
Last reviewed 2026-06-24. Facts are checked against primary/official sources where available; uncertain specifics are marked “Data not yet verified” rather than guessed.