A reproducible way to study public competitor ads
Advertisng treats public-ad data as evidence of what was observable, not as a substitute for an advertiser’s private performance data. This is the method behind our articles and quarterly reports.
Reviewed August 9, 2026 by Advertisng Research Team
1. Define the question before searching
A useful research task starts with a decision: identify offer patterns, compare messaging, map formats, or prepare a creative brief. We record the market, platform, country, date, search terms, and filters so another researcher can repeat the work.
2. Collect public evidence
Advertisng currently supports research across public Meta and Google ad sources. We preserve the source platform, advertiser identity, visible creative, public activity state, and collection time when available.
3. Classify observable attributes
- • Offer and call-to-action language
- • Creative format and visible production pattern
- • Hook, proof device, objection, and audience cue expressed in the ad
- • Repetition, variation, and longevity observed within the collection window
4. Separate observation from inference
“The advertiser ran three variants” is an observation. “This angle converted best” is not supported by a public library alone. Researchers may form hypotheses, but briefs must label them as hypotheses to test against first-party campaign data.
5. Apply quality controls
- • Deduplicate records before counting.
- • Do not treat active status as proof of profitability.
- • Verify platform or vendor claims on official pages on the review date.
- • Record collection gaps, unavailable media, geography limits, and classification uncertainty.
- • Use small samples for worked examples unless the full corpus was queried reproducibly.
6. Convert findings into decisions
The output is a short evidence table, a list of hypotheses, and a brief with explicit constraints. Original creative work should adapt the underlying customer insight—not copy a competitor’s protected expression.