How does MetaTraffic deal with users who have cookies disabled? I noticed in class.asp that Session.SessionID is used to track unique users. For a user with cookies disabled, this would track them as a unique user on every page.
Submitted by bjrcreations on Sat, 09/27/2003 - 01:40.
That would mean that every link on the page has to append the SessionID as a QueryString. I don't think that's how MetaTraffic is doing it. I guess I want to know because I am interested in the reliability of the unique users numbers that are generated.
Appending the session id to the url in a "hidden" way is proper to php, it doesn't apply to ASP.
MetaTraffic is using the SessionID provided by IIS (the server), and yes, cookies must be enabled to get a SessionID, and if I'm not mistaken, that applies to php as well.
Let's face the fact, all sites using shopping carts are emphasing on the fact that cookies must be enabled to use their facility, simply because they're using the sessionid as part of the identification of the client during the shopping session.
Would it be MetaTraffic or any similar product, determining the "unique users" based on the IP address, the SessionID or a combination of them is simply not sufficient nor is it a realiable way to identify the distinct users.
I did some tests on IIS4 and IIS5 based servers, and it happened that the same sessionid was taken by two consecutive users, or even the same user reconnecting with a different IP because his initial connection dropped, but was given the same SessionID as it had before the disconnect. Obviously there is some other internal server validation which is unknown to us, or not accessible programmatically.
For the cherry on the sundae, the AOL proxies are working in such a way that the user's IP address is changing dynamically at rapid intervals. For instance, one day I had 8 active users as per MetaTraffic. In fact, this was the same user for which the IP address was changed 7 times within 2 minutes. As if this wasn't enough, the IIS server went ahead and created a new sessionID for each IP ; not quite nice. So globally, AOL is screwing the things up royally. Thank God I didn't have more AOL users connected at the same time!
Sincerly, if that would not be of potentially impacting the users, I would definitively consider blocking any access to my site from AOL related and crappy non-sense.
I'm actually testing another approach to identify unique users, although it might not be perfect, but might lead to more accurate figures. My plans are to include the browser's string, the supported languages string along with the IP and sessionID in the process of determining unique users.
This will be solely based on the fact the the browser's string and the supported languages string pairs would not be duplicated (used by 2 different computers that is), within a given interval, normally covered by the timeout setting. So if the IP/Session pair values is changing but the Browser/Languages one isn't, always within a pre-established time frame, the odds are that we'll be dealing with the same client's computer, even if the IP/SessionID changed down the road. Does it make sense ?
This is only an hypothesis, but I'll write some code and perform an analysis of the records I have in my DB, then compare it with the results returned by MT. If I'm way off, obviously I'll have to revise the approach, on the other hand, if I'm rather close, I might be able to consider it as a potentially viable method.
Nevertheless, I'll probably include my code as an extra in my scripts, and show both MT and my values in the reports, just for the sake of comparaisons and study.
Time will tell, but as of now, I don't know of any product providing 100% accurate stats, and more specifically when dealing with unique users. Whoever can come up with a 100% accurate algorithm, the beer s on me! :D
The page you proposed, is explaing/exploring few possibilities to determine "unique users".
Globally, they mentionning elements I covered in my previous post, using a different wording. The result is pretty much the same. There is no way to determine with precision the "unique users", among some other stats that may "suffer" from information availability limitations.
In that process, there are too many "IF" to come up with a precise answer. I suggested an approach, which may be somehow viable and other people will come up with something else.
Let's be honest, if there would not have been abuses on cookie usage by a lot of crappy adrvertising firms, the odds are that people would be more willing to enable them, allowing developpers to benifit from the accuracy they can provide for some processes.
I think there is a missunderstanding on the users end as to what the cookies really are, thinking that this is something that will jeopardize their privacy.
Crucial informations, such as the e-mail addresses, are not available via code, and this is good. However there are other pieces of information, mainly for statistical purposes, that can be retrieved and used.
Regardless of what we say, I don't think we'll be able to change people habits today. We'll have to deal with what we have, and perhaps provide some information to the clients about the benifits of using cookies or other fonctionalities, such as JavaScript.
Users with cookies disabled?
I am not exactly sure, but I think that the sessionid is propagated through the URL. I know that PHP compensates by putting it in the URL 'invisibly'.
Users with cookies disabled?
That would mean that every link on the page has to append the SessionID as a QueryString. I don't think that's how MetaTraffic is doing it. I guess I want to know because I am interested in the reliability of the unique users numbers that are generated.
Users with cookies disabled?
Good day,
Appending the session id to the url in a "hidden" way is proper to php, it doesn't apply to ASP.
MetaTraffic is using the SessionID provided by IIS (the server), and yes, cookies must be enabled to get a SessionID, and if I'm not mistaken, that applies to php as well.
Let's face the fact, all sites using shopping carts are emphasing on the fact that cookies must be enabled to use their facility, simply because they're using the sessionid as part of the identification of the client during the shopping session.
Would it be MetaTraffic or any similar product, determining the "unique users" based on the IP address, the SessionID or a combination of them is simply not sufficient nor is it a realiable way to identify the distinct users.
I did some tests on IIS4 and IIS5 based servers, and it happened that the same sessionid was taken by two consecutive users, or even the same user reconnecting with a different IP because his initial connection dropped, but was given the same SessionID as it had before the disconnect. Obviously there is some other internal server validation which is unknown to us, or not accessible programmatically.
For the cherry on the sundae, the AOL proxies are working in such a way that the user's IP address is changing dynamically at rapid intervals. For instance, one day I had 8 active users as per MetaTraffic. In fact, this was the same user for which the IP address was changed 7 times within 2 minutes. As if this wasn't enough, the IIS server went ahead and created a new sessionID for each IP ; not quite nice. So globally, AOL is screwing the things up royally. Thank God I didn't have more AOL users connected at the same time!
Sincerly, if that would not be of potentially impacting the users, I would definitively consider blocking any access to my site from AOL related and crappy non-sense.
I'm actually testing another approach to identify unique users, although it might not be perfect, but might lead to more accurate figures. My plans are to include the browser's string, the supported languages string along with the IP and sessionID in the process of determining unique users.
This will be solely based on the fact the the browser's string and the supported languages string pairs would not be duplicated (used by 2 different computers that is), within a given interval, normally covered by the timeout setting. So if the IP/Session pair values is changing but the Browser/Languages one isn't, always within a pre-established time frame, the odds are that we'll be dealing with the same client's computer, even if the IP/SessionID changed down the road. Does it make sense ?
This is only an hypothesis, but I'll write some code and perform an analysis of the records I have in my DB, then compare it with the results returned by MT. If I'm way off, obviously I'll have to revise the approach, on the other hand, if I'm rather close, I might be able to consider it as a potentially viable method.
Nevertheless, I'll probably include my code as an extra in my scripts, and show both MT and my values in the reports, just for the sake of comparaisons and study.
Time will tell, but as of now, I don't know of any product providing 100% accurate stats, and more specifically when dealing with unique users. Whoever can come up with a 100% accurate algorithm, the beer s on me! :D
Regards.
Users with cookies disabled?
http://www.netiq.com/support/log/understandlog.asp
This article might help you out in your new coding endeavors :)
Cheers.
Users with cookies disabled?
Good day,
The page you proposed, is explaing/exploring few possibilities to determine "unique users".
Globally, they mentionning elements I covered in my previous post, using a different wording. The result is pretty much the same. There is no way to determine with precision the "unique users", among some other stats that may "suffer" from information availability limitations.
In that process, there are too many "IF" to come up with a precise answer. I suggested an approach, which may be somehow viable and other people will come up with something else.
Let's be honest, if there would not have been abuses on cookie usage by a lot of crappy adrvertising firms, the odds are that people would be more willing to enable them, allowing developpers to benifit from the accuracy they can provide for some processes.
I think there is a missunderstanding on the users end as to what the cookies really are, thinking that this is something that will jeopardize their privacy.
Crucial informations, such as the e-mail addresses, are not available via code, and this is good. However there are other pieces of information, mainly for statistical purposes, that can be retrieved and used.
Regardless of what we say, I don't think we'll be able to change people habits today. We'll have to deal with what we have, and perhaps provide some information to the clients about the benifits of using cookies or other fonctionalities, such as JavaScript.
Regards.