What Is Ad Intelligence? Capture, Attribution and Evidence Limits
Follow the ad-intelligence evidence chain, define its counting units and distinguish observations from performance, targeting and compliance claims.

Ad intelligence is the systematic collection and analysis of public advertising evidence: creative, observed placement, likely advertiser, destination and change over time. It helps you investigate a market and form better questions. It does not reveal a competitor's profit merely because an ad remains visible.
A useful record lets another researcher distinguish what was captured from what was inferred. That distinction matters more than a large count of screenshots.
The evidence chain#
A practical pipeline connects five stages:
Public placement → creative capture → source classification → advertiser resolution → independent destination evidence
Each stage should preserve its own timestamp and uncertainty. A network name derived from a known delivery endpoint has a different basis from a guessed label in a screenshot. A brand on a destination page may be the product owner without being the person who paid for the advertisement.
| Stage | Evidence worth keeping | Limit to state |
|---|---|---|
| Placement | Publisher or source, collection method, observation time, region/device context | One collection context is not the advertiser's entire targeting setup |
| Creative | Exact captured image or video, headline and asset identifier | The asset may be one rotation or variant |
| Classification | Network/serving signals and the basis for assigning them | Shared infrastructure and intermediaries can make attribution ambiguous |
| Advertiser | Displayed name, resolved domain and supporting connection | Brand, agency, affiliate and payer can be different entities |
| Destination | Independently resolved URL, trace/capture time and available page evidence | Geography, device, time and consent can change the destination |
OpenAdLibrary's collection is API-first where supported, with browser collection as a fallback. Public network delivery APIs are not the same as a universal searchable archive, but they are APIs: it would be wrong to say that no API exposes native placements. Collection coverage also varies by source and time.
Independent destination tracing is intended to avoid clicking a live ad. Preserve failed or incomplete traces as such; do not fill gaps by claiming that every click or every destination was followed. For more detail, see the native ad supply chain and identifying a serving network.
Define the counting units#
Before comparing datasets, write down the unit:
- A creative is a deduplicated creative record under the archive's declared matching rules.
- An observation records a capture in a particular context and time; many observations may refer to one creative.
- An advertiser is a resolved identity under the archive's rules, not necessarily a verified legal payer.
- A landing capture records a destination observation; it is not necessarily a unique landing page or a unique campaign.
Changing a headline, merging domains or changing deduplication rules can alter totals without a corresponding change in advertising activity. Cumulative records, currently visible results and estimated active ads must not be compared as the same quantity.
A reproducible report names a collection interval, cutoff, filters, attribution rules, deduplication method and missing coverage. It saves the query and explains whether totals are exact, approximate or capped. This guide does not reuse the earlier mixed June totals as a current snapshot.
Read historical examples critically#
The following images remain from the June 2026 guide as historical evidence of messages observed, not endorsements of their claims or current availability.

This example illustrates a deadline and tax-related claim. The capture does not substantiate the advertised tax outcome or prove that its wording converted.

The image and message can be studied as creative choices. Their presence does not demonstrate product efficacy or profitable customer acquisition.

An observed pet-content message can prompt a question about the destination's offer. An observation span is not proof of uninterrupted buying.

This is a captured health claim requiring independent substantiation. Its appearance in an archive provides no medical validation.
Longevity is an observation, not a profit metric#
First seen and last seen bound observations in your sample. They do not prove that an ad ran continuously between those times, or that it started and stopped at those moments. Gaps can reflect collection coverage, rotation, targeting or unavailable inventory.
Repeated observations can make an ad worth investigating. They cannot disclose spend, click-through rate, conversion rate or profit. An advertiser might continue a losing campaign, pursue a brand objective, have delayed revenue, buy only small volumes, or fail to notice a problem. A creative disappearing can reflect a completed promotion, a budget change or a collection gap rather than a failed test.
Likewise, a newly observed country is evidence of a new observation in that country. Calling it deliberate scaling requires more support. Keep the hypothesis separate from the fact.
Ad transparency and the DSA#
Transparency provides disclosures or public evidence. Intelligence adds structured analysis; it should not turn disclosure gaps into invented certainty.
DSA Article 39 requires designated very large online platforms and search engines presenting advertisements to maintain a searchable repository, including API access, for the period an ad is presented and one year afterward. It specifies fields such as ad content, the party on whose behalf it is presented, a different payer where applicable, timing and targeting information. Scope and application depend on the regulation and designation provisions; this is not a rule that every website maintains a complete archive of every ad.
The Commission's July 1, 2025 transparency-reporting change concerns harmonized content-moderation reporting. That date is not the general start date for Article 39 ad repositories. Content-moderation reports, the DSA Transparency Database and ad repositories serve different purposes.
Public visibility is also not blanket legal permission for every collection or reuse. Consider the applicable access conditions, privacy, copyright and other rules for the actual activity. An archive is evidence to assess, not a legal clearance certificate.
Turn a capture into a useful decision#
For a media buyer, the useful output may be a brief that identifies a message, the promised outcome, the destination structure and a question to test with truthful original creative. For a brand team, it may be an evidence packet distinguishing a suspected unauthorized use from a confirmed one. For a researcher, it may be a time-bounded comparison with explicit coverage limits.
A good packet includes the underlying records, method and unresolved questions. For example: “We observed this message on these dates; its destination emphasized a trial; we will test whether explaining our own trial more clearly changes qualified signups.” That is defensible. “The competitor has found a winner, so copying it will work” is not.
Use your own first-party results to evaluate performance. External ad evidence can inform a hypothesis, but it cannot replace the revenue, cost, consent and attribution records needed to judge your campaign.






