Use Cases

What is Ad Verification?

Ad verification is the practice of checking that ads ran as they were bought: the right creative, the right geography, an acceptable page, a viewable position. Because ad delivery is decided from the request's IP, device and profile, verification teams load publisher pages from many locations and record exactly what was served.

Verification bundles several separate checks. Creative correctness asks whether the served asset is the one that was bought and whether it renders inside its slot. Targeting compliance asks whether a campaign bought for Germany is showing in Germany. Brand safety examines the page the ad landed on, its content category and what sits next to the placement. Fraud checks look for invalid traffic, spoofed domains, stacked or one-pixel iframes, and click-through chains that redirect to somewhere the advertiser never approved.

Proxies are involved because delivery is decided from the request itself. The auction reads the country, region and ASN of the IP, the device type from the user agent and client hints, the time of day, and whatever audience data attaches to the profile. Requesting a Brazilian publisher from a Frankfurt cloud instance returns whatever fills Frankfurt inventory. Bidders also filter known hosting ranges, so checks from datacenter addresses often collect house ads, blank slots or a fallback creative, an outcome produced by the test setup rather than the campaign.

The mechanics are browser work. A browser loads the publisher page, records a HAR of every request in the chain, screenshots the viewport, and follows the click-through to its final URL to catch redirect abuse. Viewability needs real layout, so the browser has to render at a realistic viewport size and scroll the page. Verification traffic is itself automated traffic and can be counted as invalid by the same systems it inspects, which is why measurement vendors register their crawlers rather than hiding them.

Sampling design decides how useful the output is. Several cities per country, mobile alongside desktop, different hours, and repetition across days, because inventory rotates and one check proves very little. Keep the artefacts, meaning the HAR file, the screenshot, the timestamp and the exit IP with its geography, since that bundle is the evidence in a makegood conversation with a publisher or network. Discrepancies between publisher-counted and advertiser-counted impressions usually trace back to blocked pixels, viewport differences or traffic filtered as invalid.

Where you meet it

You come to this when a client asks why their Spanish campaign appeared in a screenshot taken in Mexico, or when a brand safety report lists a domain nobody bought. Answering means reproducing the impression from the kind of connection the real user had, with the right country, network type and device profile. Cloud addresses answer a different question, because much of the demand chain classifies them before the auction even runs.

Common questions

Why can't I verify ads from a cloud server?

Much of the programmatic chain classifies known hosting ranges as non-human traffic and either withholds bids or serves fallback inventory, so you record house ads and blanks instead of the campaign. Requests from residential or mobile addresses in the target market are treated as ordinary users, and the creative that renders is the one a person would see.

What is the difference between ad verification and ad fraud detection?

Verification is the wider activity, asking whether the ad ran where, when and how it was bought, beside acceptable content and in a viewable position. Fraud detection is the subset concerned with traffic and inventory that was never legitimate, such as bots, spoofed domains and hidden placements. Most verification programmes report both from a single collection run.

How many locations do I need to check a campaign?

Enough to match the buy. A national campaign usually needs several cities rather than one, since delivery differs by region and by network, and mobile inventory has to be checked from mobile connections. Repeating checks over time is worth more than adding locations, because creative and inventory rotation change what any single check records.

Related terms

Geotargeting
Geotargeting is choosing which country, region, city or carrier your proxy traffic exits from, so a site sees a visitor located where you want. You select it through parameters in the proxy username, a country-specific hostname, or a dedicated port, and the site reads that exit IP against whichever geolocation database it uses.
Web Scraping
Web scraping is the automated collection of data from websites. A program requests pages over HTTP, extracts fields from the HTML or JSON that comes back, and writes them as structured records. Scale ranges from one script polling a page to distributed crawlers that rotate proxies and render pages in a browser.
Multi-Accounting
Multi-accounting is operating several accounts on one platform from the same machine or team. Platforms link accounts by IP address, cookies, device fingerprint and behaviour, so each identity needs its own consistent set of those signals, typically a separate antidetect browser profile paired with a separate and stable proxy exit.
Account Farming
Account farming is registering accounts in bulk and warming them with ordinary activity before they are used. New registrations carry the least trust on any platform, so operators age them and build a history, keeping each one on a consistent IP, device fingerprint and behaviour pattern. Most platforms prohibit the practice in their terms.
Price Monitoring
Price monitoring is the repeated collection of competitor prices, stock and promotions from retail sites and marketplaces. Retailers vary what they display by country, currency, delivery postcode, device and account, so the IP and session of the collecting client decide which price it records. That is why the checks run through geo-targeted proxies.
SEO Monitoring
SEO monitoring is the continuous tracking of where a site ranks in search engines for a defined set of queries, broken down by country, city, language and device. Because search results are localised and often personalised, rank trackers run each query from an IP address in the target market rather than from one office connection.
SERP
A SERP is a search engine results page, meaning everything an engine returns for one query. Alongside organic links it can carry ads, a local map pack, People Also Ask, images, video, shopping units and an AI-generated summary. What appears changes with the query, the searcher's location, language and device, so two SERPs for the same words are rarely identical.

Real 4G/5G mobile and residential IPs

PROXIES.SX runs carrier IPs in 100+ countries with HTTP and SOCKS5 on every endpoint. $4/GB down to $2.40/GB at volume, free endpoints and rotation, and your GB never expire.