Forum Moderators: DixonJones
Why don't you try an offline weblog analyzer? There is software (some free ones, some commercial) which you can install on your own PC. I am working with the free WebLog Expert Lite. I am simply copying the Logs via FTP to my local PC and direct the the SW to the local path where my logs are. It is basic, but gives me some ideas.
Benni
P.S. I am still little confused about the log file format with my two different webhoster. One is having e.g. mydomain1111017600.gz AND referrer_log.1111017600.gz, the other one is just having e.g. access_log_2005-03-07.gz which gets me confused, but the SW apperently is fine with it. If anyone has inputs on that one, I would be thankful
Benni
It is completely free and it'll go out and get the log files for you so you don't have to manually get them yourself.
I've been using it since it was release a few days ago and I've been blown away.
If you want a link just send me a sticky mail and I'll send it over.
-Jason
Awstats suggests I got 1700 visitors, whilst Weblog Expert Lite says I got 2500.. Interesting! Will see what the conversion rates are like.. Can anyone explain the difference though?
Exactly why I don't use log files [webmasterworld.com]
The issue of differences between results for different stats programs has been discussed here before and I'll try to find some of those old items because they had some very good posts.
Some programs filter out spiders and bots; others don't. Spiders and bots can be up to 50% of your traffic.
Some programs use semi-sophisticated ways of tracking visitors such as a persistent cookie. Others try to use a session cookie, mistakenly. Others use a concatenation of the IP address and the User Agent field. Others just use the IP address. The difference between visitor counts, for a busy site, using the first vs. last method can easily be 2x.
Some programs that try to use cookies will be clumsy about matching the first cookieless hit to the rest of the visit, and can count those first hits as separate visitors.
Some programs don't ignore images and so will count as visitors any orphan hits to images, such as images displayed in an e-mail or on another site.
It's also possible that you are misreading their output and you are accidentally comparing visit statistics with visitor statistics.
And so on. Your two programs aren't wrong, they are just using different methods and assumptions, and they're too low-level to either tell you the possible differences or let you change how they work. What they do is not magic. You can do pretty much the same thing with Excel and an investment of some of your time, in fact everybody do that at least once in their life. It will help you realize how different programs come up with different results.
Regarding the different log file names: these are just different ways Apache servers store and name logs, and one method is older than the other. They may or may not contain exactly the same kinds of information, depending on how the hosting people bothered to set up logging. Open them up in a text editor or Excel (they are space delimited) and puzzle out the contents. Search for articles on "Apache log file formats" for help in knowing what the fields are. Even primitive log analyzers can handle the different formats as they are among the most common formats in the world.
Let's please not blame everything on log files all the time. That's getting old. We could try answering the question.The issue of differences between results for different stats programs has been discussed here before and I'll try to find some of those old items because they had some very good posts.
Some programs filter out spiders and bots; others don't. Spiders and bots can be up to 50% of your traffic.
Some programs use semi-sophisticated ways of tracking visitors such as a persistent cookie. Others try to use a session cookie, mistakenly. Others use a concatenation of the IP address and the User Agent field. Others just use the IP address. The difference between visitor counts, for a busy site, using the first vs. last method can easily be 2x.
Some programs that try to use cookies will be clumsy about matching the first cookieless hit to the rest of the visit, and can count those first hits as separate visitors.
Some programs don't ignore images and so will count as visitors any orphan hits to images, such as images displayed in an e-mail or on another site.
It's also possible that you are misreading their output and you are accidentally comparing visit statistics with visitor statistics.
And so on. Your two programs aren't wrong, they are just using different methods and assumptions, and they're too low-level to either tell you the possible differences or let you change how they work. What they do is not magic. You can do pretty much the same thing with Excel and an investment of some of your time, in fact everybody do that at least once in their life. It will help you realize how different programs come up with different results.
Regarding the different log file names: these are just different ways Apache servers store and name logs, and one method is older than the other. They may or may not contain exactly the same kinds of information, depending on how the hosting people bothered to set up logging. Open them up in a text editor or Excel (they are space delimited) and puzzle out the contents. Search for articles on "Apache log file formats" for help in knowing what the fields are. Even primitive log analyzers can handle the different formats as they are among the most common formats in the world.
That was a very informative post. Thank You!
Benni_203
You must have missed the link in Receptional's post. If you check it, you will discover that he is quite knowledgeable about visitor tracking, and did, in fact, "answer the question" as you put it, quite some time ago. In fact, he "answered the question" before most of the posters in this thead had ever heard of WebmasterWorld :)
2. Just saying "exactly why I don't use log files" does not help the user. A question was asked and an appropriate response would be, well, an answer, not a self-oriented editorial remark. The responder was capable of an effective answer but instead chose to make a sweeping editorial remark.
3. The old item started by equating log files to the absence of cookies, which was outdated and wrong long before the old item was written.
4. The old item did not answer the user's question even indirectly, much less correctly or helpfully.
This is getting off the topic of the original post.
simply because they use a different algo to interpret the logs.
I personally don't use server side tracking because it is adding even more overhead to the browser load time for indifinitive data.
cookies are disabled java is disabled ect ect
so i trust server logs more - not for tracking users but for tracking page requests. which pages are more active - search engine queries can give you a good idea what users are looking for.
I use analog - it is a free program that you can use on your home pc. It is a HUGE learning curve to set up but not impossible. What i like about it is that it is totally customisable. I can glean whatever reports I want (once I set it up)
The only way you can use any tracking sofware is to consistantly use the same method and observe trends.
It is not totally accurate info but trends can be seen (provided you are using the same sofware to measure those trends).
Once you switch to a different system you can't compare it with your old one - you have to start over.
So pick one you like - hopefully that shows some detail and use it over a period of time before you can make any conclusions, don't think about 'how much traffic' think more along the lines of 'traffic going up traffic going down' traffic from bots vs from browsers.
Time spent on site by viewers (going up or down)
cached pages ect ect.
That is the only reason my server does logging. For me, how many people come to my site is irrelavent, I am more concerned about some Bonehead trying to hack/attack my server.
I do visitor logging/analysis the old fashioned way, every hit is sent to a .txt file, at the end of each day (or when I feel like it), I take the days logs, and put them into another file and archive it for myself and the Security Admin. (This procces takes about 3-5 min depending on how busy the server was.)
Reading these made me want to put in my 2 cents, but that is what forums are for, right?
Anyways, y'all have a good day...
-- The DEVI8OR
