A ghost ad test shows a neutral placeholder ad to the control group instead of your real campaign. Both groups receive the same number of ad impressions, which means any difference in outcomes is caused by your ad message rather than by the mere presence of advertising.
Standard holdout tests exclude the control group from advertising entirely. Ghost ad tests go further. They give the control group an ad that looks like yours in terms of placement and frequency but carries no campaign message. This is also called a placebo ad.
Why ghost ads are more precise
Seeing an ad has several effects beyond delivering a message. Reaching an auction bid can affect what competitors pay. Ad exposure can change browsing behaviour even for neutral content. By showing the control group a real ad impression with a neutral message, you isolate the pure persuasive effect of your creative.
Standard holdouts measure: the effect of advertising versus no advertising. Ghost ad tests measure: the effect of your specific message versus a neutral presence. For creative testing, the ghost ad approach is considerably more precise.
How a ghost ad is structured
- Same ad format and placement as the real campaign
- Same targeting parameters and bid strategy
- Neutral creative content, typically a public service message or a non-competing brand
- Same exposure frequency and timing as the test campaign
When to use a ghost ad test
Ghost ad tests are most useful when you are trying to isolate the creative effect of a specific campaign rather than the value of a channel overall. If you want to know whether a new brand campaign is changing purchase intent, a ghost ad test separates the messaging effect from the noise of simply appearing in front of your audience.
They are also useful for measuring awareness and brand recall, where a pure holdout test would confound the effect of seeing any ad with the effect of seeing your specific message.
Ghost ad tests are operationally more complex than standard holdouts. You need a ghost creative ready to run, and the platform must support running two different creatives to split audiences while maintaining equal reach to both groups.
Ghost ad tests on social platforms
Meta previously offered a ghost ad testing feature through its Conversion Lift tool, where a neutral public service announcement served as the control creative. This approach has become less available as platform privacy constraints have tightened, but the methodology can still be replicated manually with careful campaign setup.
On programmatic platforms, ghost ad tests can be designed by a third-party ad server using audience splits and two campaign line items: one for the real creative and one for the ghost.
Interpreting ghost ad results
A ghost ad test produces a lift estimate that reflects the incremental value of your message above a neutral baseline. This is a tighter, more demanding test than a standard holdout. Expect lift to be lower than in a standard holdout test, because you have already removed the effect of simply being present in the ad environment.
If your ghost ad lift is zero, your creative is not persuading anyone beyond what a neutral ad presence achieves. That is a signal to review the message, not the channel.
Is a ghost ad test the same as an A/B test?
They are related but different. An A/B test typically compares two versions of your creative to find which performs better. A ghost ad test compares your real creative to a neutral control to measure its absolute persuasive effect. Ghost ad tests are about causality; A/B tests are about optimisation.
What qualifies as a good ghost ad creative?
An effective ghost creative must be genuinely neutral: it should not promote any product or create brand associations. Public service announcements from charities are commonly used. Avoid ghost ads that could accidentally help a competitor or create negative brand associations in either group.
Do ghost ad tests work for video advertising?
Yes. For video, the ghost creative should match the format and length of your real ad but contain neutral content. The same targeting and delivery parameters should apply to both groups. Measurement of video ghost ad lift typically focuses on brand recall, awareness, or search uplift rather than direct conversion.