How Does Ad Verification Work? Checks, Evidence, and Limits

Ad verification checks whether an advertisement appeared in the context a campaign intended. Teams compare observations about placement, creative, location, visibility, and traffic with campaign settings and delivery reports. A screenshot can document what loaded during one visit; it cannot, by itself, prove that an impression was billable, a visitor was human, or a publisher committed fraud.

A useful verification process starts with a question you can test, captures the conditions of the test, and separates an unexpected result from its possible causes. Here is how to build that process without treating every mismatch as evidence of wrongdoing.

What does ad verification check?

Different checks answer different questions. Keep their evidence separate so a successful page visit does not become a blanket statement about campaign quality.

• Placement and context: Did the creative appear on the expected page, app, or section? Was the surrounding content consistent with the campaign's requirements?
• Creative and destination: Did the right version render, and did its destination match the intended landing page?
• Geographic delivery: What appeared under a documented location and browsing setup?
• Viewability: Was the impression measured as having an opportunity to be seen under the applicable measurement method?
• Traffic quality: Did platform reports or specialized measurement identify invalid activity that needs investigation?

Visual QA is strongest for the first two questions. Viewability and invalid-traffic assessments require appropriate measurement and campaign evidence; loading a page through a proxy does not provide those measurements automatically.

Define the expected result before collecting evidence

Write down the campaign, placement, creative version, destination URL, target geography, device class, and test window. Identify the setting or agreement behind each expectation. Without this baseline, a different creative may look like an error even when an experiment, frequency cap, or schedule explains it.

Choose a test you are authorized to run. For example: confirm that an approved landing-page variant loads from a specified market during a scheduled campaign. Avoid generating artificial clicks or repeated interactions that could distort reporting or billing.

Capture a reproducible observation

Record the page URL, timestamp with time zone, device and browser, session state, network location, and observed result. Include a screenshot when it helps document rendering, but retain enough context to understand what that image actually represents.

Cookies, consent choices, login state, language, auction conditions, inventory availability, and targeting rules can all change what appears. A missing ad in one browser session is an observation to investigate, not proof that a campaign failed to deliver.

For a multi-step journey, preserve the relevant session conditions. If the test intentionally changes location or identity, record that change rather than comparing the visits as though they were identical.

Compare observations with campaign reports

Match the test window and scope to the reporting source. A visit to one placement cannot explain aggregate performance across an entire campaign. Check whether the impression, click, or conversion being discussed is actually represented in the reporting system.

Google describes invalid clicks as activity that does not reflect genuine user interest and explains how it handles detected invalid activity in its Google Ads invalid-click guidance. That platform process is separate from your own screenshot review. Use the platform's definitions when interpreting its metrics.

Where residential proxies fit

A residential proxy can provide an additional network vantage point for an authorized location-based check. It can help a team compare what loads through different egress locations while keeping the rest of a test documented.

A proxy does not guarantee the same experience as every resident of a country. It also does not guarantee invisibility, human traffic, viewability, or fraud detection. Geography is one input among several, and the target may apply its own access restrictions.

See the ad verification proxy guide for infrastructure considerations. If a check spans several requests, the distinction between sticky and rotating proxy sessions helps explain which conditions should remain stable.

Investigate mismatches before escalating

1. Confirm the baseline. Recheck the campaign settings, schedule, and creative version.
2. Repeat a bounded test. Keep the device, session, and test conditions consistent where possible.
3. Change one variable. Compare a different authorized location or browser only after documenting the original result.
4. Correlate evidence. Review delivery reports and relevant measurement alongside the observation.
5. Escalate a specific discrepancy. Share the expected result, actual result, timestamps, and reproducible steps.

Label a finding as unresolved when the evidence cannot distinguish a configuration issue from an auction outcome or technical failure. This is more useful than assigning a fraud explanation prematurely.

Turn verification into a repeatable QA process

Keep a short record for every check: the question, expected behavior, test conditions, evidence, interpretation, owner, and next action. Separate confirmed configuration defects from hypotheses that still need investigation.

Magnetic Proxy's ad verification use case describes how proxy access can support this workflow. Start with the observation you need to make, then select the access and session setup that supports that test.