On 9/26/01 7:14 PM, "John R Levine" <[EMAIL PROTECTED]> wrote:


> It's only a web bug if the url is unique to the recipient.

Unless, of course, they're back-tracking through the referrer to your IP.
If, for instance, they track your visits to the web site, they can do a lot
of user tracking just by watching what other hits come from those IP ranges.
There are obvious limitations to that, but you can get a lot of useful data
that way. 

Or maybe they're tracking back on your path geographically (think Akamai or
colo-ed round-robin stuff).

Or maybe they're using user-ID encoding in their hostname schemes to hide
the user tracking you're looking for in the URL part.

Or maybe they aren't doing any of that. But there are lots of techniques you
can use to track back lots of information that aren't obvious, but may well
be there. Just because there aren't obvious bugs doesn't mean they arne't
tracking.

Oh, and while I'm flapping my gums, just because there ARE bugs, don't
assume they ARE user tracking. Those bugs are also quite useful for
aggregated data that may or may not be tracked back to an individual. They
may not care about individual actions, but are using bugs to track
statistical samples or aggregated groupings of some sort. Not that I'd know
anything about that, you understand.

> If you grant that formatted mail is permissable at all (personally, I'd
> have rather gotten that message in plain text), I don't see how references
> to generic images make it any worse.

That's why I think it's important you keep plain-text as an option. On the
other hand: given equal choice, 80+% now prefer an HTML version of our
lists. That's not data based on an assumed default choice, either.

> On the third hand, I've heard of one moderately legit use for web bugs:
> people sign up for a list, you send them multipart/alternative with a web
> bug, and if there's a hit on the web bug you know they have a mail program
> that renders HTML so you can stop sending them the non-HTML version.

There are lots of legitimate uses.

To name one other: say you publish a newsletter. Say you want to know
whether people are interested in what you're publishing. Bug the URLs, and
see which ones they select. That gives you great data on which pieces the
audience use, and which they don't -- without the time and expense of
running surveys, which may or may not be statistically valid. You can get
real-time feedback on what stuff you pbulish they like; so you can publish
more of that, and less of what they don't care about.



Reply via email to