Full Answer

Redirects fall into two camps: those that preserve the query string and those that quietly drop it. The dangerous ones rebuild the target URL from scratch. A server-level HTTP-to-HTTPS rule, a host-canonicalisation redirect between www and non-www, or a trailing-slash normalisation will all forward the visitor but leave utm_source and utm_campaign behind unless the rule is written to carry parameters.

Geography and device add more hops. Regional redirects that send a visitor to a country subdomain, and mobile redirects to an m. subdomain, are classic UTM killers because the redirect logic usually only cares about the path. Link shorteners and vanity URLs are worse: they often resolve to a clean canonical URL with no parameters at all, so the campaign data never had a chance.

Email is its own trap. Many platforms wrap every link in a click-tracking redirect that replaces your UTMs with theirs, so the visit lands tagged as the email tool rather than your campaign. The practical defence is to test the full redirect chain end to end, configure redirects to append the existing query string, and capture the landing UTMs server-side at first touch so the value is stored before the next hop can erase it.