better option than http_referer ??

[TechBoyJK]

I would like to pull stats on where a user clicked a link from.

So far, my only option that I've found is to use http_referer, but that hardly works in chrome and is otherwise unreliable.

For instance, if you use a bit.ly link, they can tell you where the click came from, etc.

How can they be doing anything but using http_referer since there's no url variables being included in the link?

 

[BoberFett]

That's pretty much it. There's no foolproof method, it all depends on the agent and what it provides.

 

[mrjminer]

That is the only standardized option I know of (despite it being unreliable and able to be manipulated).

They could be using some sort of ad platform, or some other non-standard header, like X_FORWARDED_FOR or whatever. Also possible the click is intercepted from whatever by JS and adds a referrer.

You could enable the option in whatever browser you are using to see the headers, or use Wireshark.

 

[DaveSimmons]

> since there's no url variables being included in the link?

Other than the referrer, the URL itself might be unique:

www.foo.com / 1234567890-A.htm = link from AnandTech

www.foo.com / 1234567890-E.htm = link from Endgadget

The server then maps them both to the same page but can track the source.

Edit: there are also cross-site cookies, "super cookies". Wireshark will show all headers which will include any cookies being sent.