Website Statistics - Help 1

[Pav1977]

A couple of small things that I'm unsure about with relation to web:

1. Definition of 'Volume of traffic' for website statistics and how to get it on for example Google Analytics

2. Is it possible to get the info of where the users were forwarded? Is this a feature on any of the stats applications?

I would really appreciate your help,

Regards

 

[feherke]

Hi

1. Definition of 'Volume of traffic' for website statistics and how to get it on for example Google Analytics

Google Analytics uses code inserted into the HTML documents. ( As far as I know. ) So it can not track script, style, image and other king of files. Complete traffic information you can get only from the web server' s log.

2. Is it possible to get the info of where the users were forwarded? Is this a feature on any of the stats applications?

What do you mean "forwarded" ? In HTTP that word is used related to the proxies. Do you care about what proxies forwarded the request & response ?

Feherke.

http://rootshell.be/~feherke/

 

[Pav1977]

1.Definition of 'Volume of traffic' - so how is this measured what is the usual way of doing it?

2.I meant how to find the 'registering sites' - where did the users find the site - for example, another linked site or a search engine.

Cheers

 

[feherke]

Hi

1.Definition of 'Volume of traffic' - so how is this measured what is the usual way of doing it?

I do not think there is or should be a definition for that expression. ( You can google for 'define:"volume of traffic"' anyway... ) That means just the total transferred size. As I wrote, use an access log analyzer to see such data.

I would mention Webalizer, Analog, AWStats or just google for "access log analyzer".

Alternatively you could extract such information from a proxy log too, but I would ignore that possibility.

2.I meant how to find the 'registering sites' - where did the users find the site - for example, another linked site or a search engine.

You mean the site referring to yours ? That information is sent by the browser to the web server in the [tt]Referer[/tt] HTTP header. It can be the real, a fake or none. Browsers, proxies and firewalls can be configured to not send or remove that information. Some security softwares removes it by default without being asked for by the user.

What referrer information reaches the web server can be logged and then analyzed by the above mentioned access log analyzers.

Note that the CLF ( Common Log Format ) does not include the [tt]Referer[/tt], so you have to configure a custom format on your web server.

Feherke.

http://rootshell.be/~feherke/

 

[Pav1977]

Thanks for the interesting answers.

1. So I understand that the volume of traffic is the amount of data downloaded in MB for example. Is this commonly presented to clients per day figures?

2.It's good to know the protocol so thanks for that but what I need is to know how I could present this to a client. How do you present it to your clients I assumed that Google Analytics would do it but I can see it on there.
Whom can I use for the site's referral then?

Thank you once again

 

[feherke]

Hi

I skip the presenting to client part. I never needed such thing.

Just some facts :
[ul]
[li]More and more people see Google becoming the big brother watching you. And Google Analytics is the more adequate tool to track your internet activity. So they will block them. It can be done with browser add-on, security software or simply the operating system's networking feature. Yes, I personally received question for help to block Google Analytics. So it in not 100% relevant.[/li]
[li]The referrer log can divulge information to designated for public use. It is not always a good idea to know from where arrived someone to a site, so many people disables it. Yes, I personally caught a guy who made a copy of one of my pages. So it in not 100% relevant.[/li]
[/ul]
So again, I would use an access log analyzer. Of course, if is possible.

Feherke.

http://rootshell.be/~feherke/

 

[Foamcow]

Granted the only 'guaranteed' way to monitor stats is via internal webserver logs, but this is often quite a task.

Google Analytics allows 'the man in the street' to do it.

Granted it's possible to block GA, but it's also possible to get a webserver to log inaccurate data by using a proxy, by spoofing a referrer or a browserstring.

The thing with stats is that you should *never* look at the numbers themselves. Instead, you should study trends.
If you do this then no matter who's blocking what or spoofing whatever you will see patterns. If you take the law of averages into consideration then approximately the same number of users will be blocking - the errors in your stats will always be consistent(ish). It's patterns that indicate whether your site is growing or dying.

The long and the short of it is that web stats should always be taken with a pinch of salt and never at face value. You need a lot of data over an extended period to see what's really going on.

<honk>*:O)</honk>

Tyres: Mine's a pint of the black stuff.
Mike: You can't drink a pint of Bovril.

 

[feherke]

Hi

Granted it's possible to block GA, but it's also possible to get a webserver to log inaccurate data by using a proxy, by spoofing a referrer or a browserstring.

Yes but the "'Volume of traffic'" asked by Pav1977, in the access log will be the real one - the one you will see in your provider's invoice.

Of course, you are right, it will not be relevant as indicator of the site's popularity.

You need a lot of data over an extended period to see what's really going on.

Yes, another advantage of access logs. If your site was up for an extended period and today is your first day when you are interested in site evolution statistics, then log analyzers will do the job, while third party analytics will show nothing.


Feherke.

http://rootshell.be/~feherke/

 

[Foamcow]

Yes, another advantage of access logs. If your site was up for an extended period and today is your first day when you are interested in site evolution statistics, then log analyzers will do the job, while third party analytics will show nothing.

Providing you kept the logs. May ISPs only store the last 7/14/30 days.


Interestingly I recently was asked to add GA to a site that previously didn't have it. To my suprise the stats were there, pre-dating when I added the GA code.
The only thing I think can explain this is that the site owner had been using AdWords extensively.



<honk>*:O)</honk>

Tyres: Mine's a pint of the black stuff.
Mike: You can't drink a pint of Bovril.

 

[feherke]

Hi

Providing you kept the logs.

That is actually the second requirement. First you must have logging turned on...

Interestingly I recently was asked to add GA to a site that previously didn't have it. To my suprise the stats were there, pre-dating when I added the GA code.

So the rumor may be right, the big brother is watching...

Feherke.

http://rootshell.be/~feherke/

 

[Pav1977]

'then log analyzers will do the job, while third party analytics will show nothing.'

Where can I find log analyzers - the site is on IIS on Windows Server 2003

Thank you for your help - I'm learning lots from this.

 

[feherke]

Hi

Where can I find log analyzers - the site is on IIS on Windows Server 2003

A gave the links to the most popular ones :

I would mention Webalizer, Analog, AWStats or just google for "access log analyzer".

Note that I do not know IIS. Although all other ( production ) webservers are able to produce CLF logs, IIS may be different.

Feherke.

http://rootshell.be/~feherke/

 

[Foamcow]

You can always do both.

Install GA and keep your logs and run them through an analyser regularly.

<honk>*:O)</honk>

Tyres: Mine's a pint of the black stuff.
Mike: You can't drink a pint of Bovril.

 

[Stretchwickster]

Just for your info regarding sites that referred people to your site, Google Analytics has a detailed section called "Traffic Sources" which gives all sorts of information regarding search engine terms used to find your site, referring sites etc.

Clive

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
"To err is human, but to really foul things up you need a computer." (Paul Ehrlich)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
To get the best answers from this forum see: faq102-5096

 

[ChrisHirst]

IIS may be different

Nope no different, IIS produces logs as all webservers should

Another log analyser is FunnelWeb



Chris.

Indifference will be the downfall of mankind, but who cares?
Woo Hoo! the cobblers kids get new shoes.
People Counting Systems

So long, and thanks for all the fish.