Verification requires being indistinguishable from an audience member
The whole point is to see the page as a targeted user sees it. Any IP the ad network can flag as belonging to a monitoring service gets served clean creative, which is precisely the outcome that hides the problem you are looking for.
Geography is the campaign, not a variable
Campaigns are bought by region. Verifying a Berlin placement from anywhere but Germany does not verify it. Country and city targeting let you check each geo the campaign actually bought.
Mobile placements need mobile IPs
In app and mobile web inventory is served to carrier addresses. A desktop residential IP can return a materially different set of ads, and for mobile first campaigns that gap is the entire result.
What to buy
Residential for web placements, mobile for in app and mobile web. Both per gigabyte, targeted to the geographies the campaign was bought in.
Common questions
Why can I not just load the ad myself?
Because you are one location, one device and one browsing history, and the ad you are served is chosen using all three. Verifying that a campaign runs correctly in a market means requesting from that market, and confirming it is not being shown beside content you excluded means doing that repeatedly rather than once.
Which pool suits ad verification?
Residential for anything where the ad decision depends on looking like an ordinary visitor, which is most of it. Mobile where the placement is inside an app or aimed at carrier traffic, because those are served to carrier addresses and a residential address will not see them. Datacenter is rarely right here, since ad systems are among the most aggressive at recognising it.
Does the address need to stay the same between requests?
For a single check, no. For following a funnel, yes, because the pages after the first will be inconsistent if the address changes underneath them. Residential sessions here hold for up to six hours, which is longer than any verification path needs.