It's marketing for the origin site. The line of thought is that the author sees significant traffic from xyz.com in the ref query string, and considers advertising or partnering with the origin site.

Honestly, it is quite useful for niche/startup sites. I have been on both ends of conversations that began from seeing these in web analytics (as someone that saw incoming traffic from a site and reached out, and as someone that received contact from a site I linked to) - and both times it ended in a mutually beneficial partnership.

I can understand the privacy argument to some degree, but it provides no more information than the standard Referer header (and if you use analytics like Simple Analytics/Plausible, it is a lot more visible).

I’m broadly anti-tracking: it’s generally against the interests of the individual.

Query string additions are commonly used to track things. You can see that lots of people don’t want that by the existence of Firefox features like “copy clean link” and Extended Tracking Protection which proactively strips some like UTM parameters.

Some sites happily participate in what I will glibly call the tracking economy. They may benefit because the recipient will see in their logs that lots of people are coming from their site, and do something that helps their site because of that.

My rejecting query strings is a simple form of protest against that system.

If you have a popular website and you add that parameter the target easily sees who sends them traffic, that could be the base of sponsorships / affiliate arrangements for example.