Forum Moderators: DixonJones
This hasn't been a reliable way to track people who bookmark your site for quite some time now. Just one visitor using FireFox running through your site will throw this figure off. Even limiting this to IE browsers may not be accurate because the user might be using a software (like an RSS reader with an IE based preview screen) that will pull your favicon.ico.
A more reliable indicator would be to track clicks to a bookmark this site link.
Anyway, my high-bookmarking site has a higher proportion of Mac users and slightly more Mozilla browsers, whereas the users of the other one are less technologically savvy.
people who habitually make thousands of bookmarks they never use are noise in the data
I should have a t-shirt made for me that announces how proud I am to be "noise in the data."
"Noise in the data" isn't pejorative, it's just a fact of life. My simple-minded point was that statistics is a game of generalities and not absolutes. The existence of exceptions (like people who don't use the bookmarks they make) doesn't imply that the statistic itself should be tossed out.
Although bookmarking could be a valuable stat to track, using a lot of these commonly available off the shelf products you're not going to see reliable results tracking favicon.ico as an indicator.
My other point was that the nature of bookmarking has changed. The importance of this user action is becoming less important. That's not just my personal opinion. When you see a lot of new webmasters coming in trying to use favicon.ico as a bookmarking metric it's a sign to me that they may be reading some outdated advice somewhere.
This is not to say their data is accurate, just that they are trying to make it more accurate.
>> My other point was that the
>> nature of bookmarking has
>> changed. The importance of
>> this user action is becoming
>> less important. That's not
>> just my personal opinion.
Let's assume for a minute that the bookmarking event can be tracked correctly, via favicon or bookmark-this-page buttons or whatever. Given that, do you mean fewer people are bookmarking? Or when they bookmark, they're doing it for new and different reasons? Or more people are bookmarking? Or the same number of people are bookmarking, but doing it more/less times per visit? Something else?
It doesn't add up.
All of the above will account for returning visitors that show no referrer. If they do tend to have referrers, then repeat visitors are probably relying on links elsewhere including on portal pages that they've personalized with their favorite links i.e. your site.
As a general rule, I think the only way of tracking repeat visitors is to use IP or persistent cookies which (of course) have their own issues.
For a metric of general site popularity I prefer to keep track of webmail referrers (indication that people are emailing others to visit the site) and how many users take advantage of the site's "tell a friend" function.
(1) present an "Add to your favorites" link on WIN IE; many scripts presume that IE is enough; actually only Windows IE allows the AddFavorite().
(2) track how many use it; when used it loads a graphic called /addfav.gif"
<script language="javascript">
function addFavorite()
{
var fav = new Image();
fav.src = "/addfav.gif";
window.external.AddFavorite(location.href, document.title);
}
if (navigator.appName == "Microsoft Internet Explorer" && navigator.appVersion >= "4.0" && navigator.userAgent.toLowerCase().indexOf("mac") < 0)
{
document.write("¦ <a href='javascript:addFavorite();'>Add to your favorites</a>");
}
</script>
