The problem being solved
A redirect is a useful mechanism and an abusable one. The same link that lets an advertiser measure a click lets an attacker borrow a reputable domain to point somewhere hostile. Review a link in the morning, change the destination in the afternoon, and nothing in the URL ever changed.
Ad platforms cannot inspect every destination at click time. So they moved the requirement onto the tracker: the next hop must be visible in the tracking URL, and it must be the hop that is actually followed.
What a transparent link looks like
The destination lives in a query parameter, in plain sight:
https://novatrack.io/c/9fk2p1?force_transparent=true&url=https://yourbrand.com/offer
Three things matter here, and each has caught out trackers that thought they were compliant.
The parameter has to be followed. Not consulted, not treated as a hint. If the tracker stores a destination on its own side and uses that instead, the URL is lying about where the click goes, and the whole guarantee collapses.
Every path has to behave the same way. A tracker cannot be transparent on one route and opaque on another. If the parameter works on one path and is ignored on a second, that second path is enough to lose certification.
Nothing foreign may be added afterwards. The tracker may append parameters of its own, a click identifier for instance, because those are the same for every customer and cannot change the destination. Anything belonging to someone else, campaign tags, affiliate identifiers, must already be inside the visible URL. Otherwise the link shown and the link served are two different things again.
Transparency without an open redirect
Here is the tension nobody mentions. Making the destination editable in the URL is exactly how you build an open redirect, the single most common way a legitimate domain ends up serving phishing.
The answer is readability rather than concealment. Because the destination is written in the link, a hijacked link is a visibly different link: the address changes in front of the reader, and the server follows exactly what is written, never a stored target of its own. That is the protection, and it holds for every link. A link owner who wants a second layer can restrict a link to their own domains, in which case anything outside that list is refused outright, with an error, never quietly redirected somewhere convenient. What the server must never do, restricted or not, is silently substitute a different target: that is the failure mode the whole rule exists to prevent.
Why it should matter to you as an advertiser
If your tracker loses its standing, the ads using it get disapproved. Not the tracker's ads: yours. That is the whole risk in one sentence, and it is why the question deserves asking before you route a budget through anyone's redirect.
Two questions are worth putting to any provider. Can I read the destination in the link you generate? And what happens if I edit that destination to point at a domain I do not own? A provider who cannot answer the second question clearly has not thought about it.