Log File Explanation

Why they are not accurate.

[ThePhoenix]

I am working on a presentation to some high level executives at my company. I need to articulate in a non technical way, how and why log files are not as accurate as Page Tagging (Java Script).

Has anyone seen a written example of this? Can anyone help me with their own way of pitching this?

[jay5r]

Log files are MORE accurate if you want a full accounting of everything that happened. If you fail to put the Javascript code on a page, your Javascript page tagging system may fail to capture that hit - especially if it's an external service like Google Analytics that doesn't have access to your log files. Think of a large corporate site with lots of people generating content and you'll understand how much can be lost if you don't implment the page tagging fully. And obviously Javascript-based tracking only tracks things that can have Javascript embeded in them - namely HTML - it can't answer questions about other types of files you might serve (images, xml, etc.)

That said, log files on their own can't answer some important questions people want to know - Flash support, screen size, etc. And their ability to track user sessions depends on how good your analytics package is, and whether you embed session cookies in your log files. That's why Javascript-based page tagging exists - to enhance (not replace) log files.

Bottom line, it's not an either or situation. You'll always have log files. Javascript-based page tagging just enhances those log files - they put additional information in them that your analytics package can use to better digest the log file. If you use an external service like Google Analytics the extra information is put in their log files, not yours - but one way or another it's being put in a log file somewhere.

[Matt Probert]

It's all nonsense. Period.

Log files, report, accurately, requests to the http server. And that is ALL.

In simple terms, most companies use a router between their LAN and the Internet. So while every work station in the company has its own IP address, this IP address is local. To the Internet, all stations go through a single router with its own, single IP address, presenting to destination web sites requests (if they get that far and don't get handled by ISP caches along the way) of requests made by a single IP address, which many misleading packages reveal as a single user.

With complex cookies you can do some tracking of run-of-the-mill users, but technicaly savvy users may disable cookies or even delete them during a session, so again results are not to be relied upon.

Matt

[gregbo]

The issue isn't whether log files are or aren't more accurate than page tagging. Log files and page tagging measure different things. Simply put, log files measure activity as seen by the web server, whereas page tagging measures activity as seen by the web client (browser/UI).
What I would do (have done) in such a situation is to determine what my clients needs are, and provide them with the solution that best meets their needs.

Read the library thread [webmasterworld.com...] for more info.