Luna Proxies For Ad Verification

Verification is a sampling problem. Placements rotate, so a single check proves almost nothing and the design of the sample determines what you can conclude.

How Many Checks Are Enough

If a slot rotates between several creatives, one observation has a good chance of landing on a compliant one. Confidence comes from repeated independent samples spread across the day, not from a single careful look. Decide the number of samples per placement before starting, because deciding afterwards turns an absence of findings into a false reassurance.

Why Datacentre Addresses Distort It

Ad platforms filter non-human traffic aggressively, and datacentre ranges are the first thing they filter. Checking from one frequently returns either nothing or a different fill than a real visitor receives, which means the sample is not measuring the population you care about. Residential exits in the target market are what make the observation representative.

Turning It Into A Metric

Express results as compliance rate per campaign per market: the share of samples matching the booked placement. That single figure is comparable across campaigns and over time, and it is far more actionable than a list of individual discrepancies. Track the redirect chain separately, since substitution there often coexists with a correct-looking creative.

Coverage Across A Campaign

Verification budget is finite, so it should follow spend. Weight sampling toward the placements and markets carrying most of the budget rather than spreading evenly across the media plan, and treat low-spend placements as periodic spot checks. Reporting compliance weighted by spend rather than by placement count gives a figure that reflects money at risk instead of one dominated by small line items.