Ad Verification From a Real-User Vantage Point
Ad-tech and brand-safety teams check placement, geo-targeting, creative, and landing pages from real smartphones in the markets they buy.
What ad verification requires that a datacenter check cannot give you
Ad verification confirms that a campaign was delivered the way it was bought: the right creative, in the right placement, to the right geography and device class, with a landing page that resolves and matches the ad. The hard part is not the checking. It is the vantage point. An ad server decides what to serve based on who it thinks is asking, so the answer you get depends entirely on the identity of the client asking.
A verification job running from a cloud region asks with a datacenter IP, a headless browser, and a fingerprint no consumer device produces. Exchanges and their invalid-traffic filters treat that profile accordingly. The usual outcomes are a house ad, an empty slot, or a different advertiser entirely, none of which tell you what a real user in that market saw.
Geo, device, and creative move together
Targeting is layered, and each layer is inferred from a different signal.
- Geography comes from the exit IP and its registry data. A proxy exit that geolocates to a hosting ASN is not the same as a subscriber IP on a national mobile carrier, and demand-side platforms filter them differently.
- Device class comes from the user agent, viewport, and client hints. Mobile inventory frequently carries different creative, different formats, and different floor prices than desktop, so a desktop check quietly verifies the wrong buy.
- Creative and landing page can be swapped server-side per session. Cloaking works by serving benign content to anything that looks like a verification client and the real page to everyone else.
Archonum supplies the vantage point: a session running in Chrome on a real consumer smartphone, drawn from 250,000+ devices across 175+ countries. The browser and the exit IP are the same handset, so carrier, geography, device class, and fingerprint agree with each other because they are genuine, not because they were assembled to match. The datacenter, residential, and real-device comparison sets out what each origin looks like to the systems doing the classifying.
What “as-delivered” evidence means in a brand-safety report
A brand-safety finding is only as strong as the record behind it. As-delivered evidence is a capture of the placement at the moment it was served, taken from a client the ad system treated as an ordinary user: the creative, the surrounding page context, the click destination, the timestamp, and the market the session originated in.
Two industry references make that record legible to the people who receive it. Viewability is measured against the MRC Viewable Ad Impression Measurement Guidelines (v2.0, August 2015), which count a display impression as viewable when at least 50% of its pixels are in view for one continuous second, and a video impression at two continuous seconds. Supply-path questions resolve against the publisher’s ads.txt file, the IAB Tech Lab standard for declaring which sellers are authorized to sell that inventory.
Screenshots alone rarely settle a dispute. What makes a report defensible is that the session it came from was ordinary: a real device, in the targeted market, on the device class the line item was bought for. That is the same argument as real devices over emulators, applied to measurement rather than collection.
Measure delivery, do not touch it
Verification is a read operation. Every impression a verification session renders is an impression someone paid for, and every click is a billable event on a real advertiser’s account. That constrains the method: request a placement once per check rather than looping it, and resolve landing pages without firing the paid click when a direct request to the destination answers the question.
Archonum is not a tool for generating ad traffic. It exists so verification teams can observe what an ad system delivered. It is not for producing impressions, clicks, or engagement on any ad platform, and traffic that inflates a metric is fraud regardless of who runs it or why.
The ad verification use case walks the workflow through end to end. If you need to check markets or device mixes your current setup cannot reach, tell us what you are trying to verify.
FAQ
Ad servers and invalid-traffic filters classify the client by IP, ASN, and browser fingerprint before deciding what to serve. A request from a hosting ASN with a headless browser profile is commonly filtered, so the slot returns a house ad, a blank, or a different advertiser than a real user in that market would have been served.
Frequently, yes. Mobile and desktop are usually transacted as separate inventory with different formats, sizes, and floor prices, and the creative is selected from the user agent, viewport, and client hints. Checking a mobile line item from a desktop browser verifies a different placement than the one that was bought.
A capture of the placement at the moment it was served, taken from a client the ad system treated as an ordinary user. It records the creative, the surrounding page context, the click destination, the timestamp, and the market the session ran in, so the finding can be reproduced rather than asserted.
No. Archonum is infrastructure for observing what an ad system delivered, not for producing traffic on it. Generating impressions or clicks to influence billing or performance metrics is ad fraud and is outside what the service is for.
Ready to get started?
Run the same script you're running today against Archonum and compare the block rate yourself.