Well, here's my first real post on Gottaquirk. Rob asked me to do a bit of a write up on the old debate around web usage measurement. Honestly I got a little freaked out by the request as I hadn't given the subject much thought till Rob brought it up and I have very little real experience. As a result this post has taken an insane amount of time to write. And rewrite. Over and over again.
I didn't have a clue what I was going to write about, but luckily for a newbie like me there is a wealth of information available on the topic. I've tried to formulate an opinion based on what I've picked up but being aware of my lack of experience has delayed my posting this. Eventually I ran out of excuses and had no option but to suck it up and put my neck on the line. So here goes:

Web analytics and conversion optimisation are becoming ever more prominent for online marketers. Currently, there are two championed ways to measure web usage; log files and JavaScript tags. A few weeks ago Avinash Kaushik wrote a great post on how the two measure up, and his views got the ball rolling on this post.

In a nutshell, part of a web server's function is to record its activities in log files. It didn't take long for marketers to start making use of this data to analyse how many page-views and visits their sites were logging up. A stumbling block for this method arose with the popularity of caching. On a repeat visit to a site the server won't necessarily recognise the request as it'll be processed through the browser's cache. Thus the user's interaction with the site won't be picked up rendering the data incomplete.

Recent times and trends have seen general favour (seemingly including Kaushik’s) moving away from the web logs to JavaScript page-tagging. In page-tagging a third-party web analytics server provides a cookie to each user which then identifies them throughout all of their visits to the site.

Each option presents its own unique set of advantages and disadvantages. Both are hotly contested, and the pros and cons of each have long been debated. Initially I was going to start out by listing these pros and cons, but ultimately that ended up more as a compilation of the more prominent bloggers' lists rather than as my own opinion.

So, here we go, my first opinion posted as an intern...
And it is ... use both page tagging and log file analysis.

Ok, so that's a bit of a cop-out on the opinion stakes, but really, if you have the money then implementing both will ensure that the minimum amount of data is lost. And with the combination resulting in any redundant information being discarded, from a usability point of view, there don't seem to be that many reasons for not doing so.

Of course the one reason that will inevitably raise its head is the cost of implementing both measures. Google Analytics is a free service utilising JavaScript, but users may want a more complete or a more tailored package which will usually cost them a monthly fee. Log file analysis normally involves a once off payment for the software.
If you will be forced to choose just one option then the main points that I'd consider would be the following:

• With JavaScript tagging all of your information will be in a third-party's hands, putting you in a disadvantageous position if you would wish to change to a new service provider.
• Conversely, JavaScript tagging gives the analytics department and the developers more focus and flexibility towards their own responsibilities.
• Log file analysis can be onerous and tedious, taking your organisation's focus away from the task at hand.

But probably the most telling argument is that in implementing page-tagging you can potentially damage the site's usability. I reckon that this should be the swinging vote in choosing between the two. If you can get away with using the richer reporting page-tagging and not compromising on your usability and conversion, then go for it. If you can't, then in doing so you'd be losing the game before you'd even walk out onto the field.

So there you have it. First option – use both log file analysis and page-tagging where the strengths of one covers the weaknesses of the other. Second option – go for page-tagging if you're comfortable with how it'll affect your site.

And leaving the best option for last – make up your own mind. Originally I had a few pages of pros and cons, many taken from Kaushik's post and from Dr Stephen Turner's musings (he in large part responsible for the Wikipedia entry on web analytics).
With so many minds contesting the benefits of each it is doubtful whether one post is going to communicate why the one option is better than the other. What can be said is that you have to find which option works best for you given your particular situation and goals.

I'm still very much in the process of weighing one up against the other and I'd be interested to hear your points of view on this debate. Even if the intention is to utilise both log files and JavaScript, identifying the benefits of one option can get the ball rolling on new ways to use and interpret the information.

I imagine that personal preference could hinge on one valid, powerfully relevant point being made, and as I'm new to this game I'm trying to hear as many points as possible. I'm tempted to say that once sufficient points have been made I would then be able to finalise my opinion, but it's clear that any argument is not going to support page-tagging or log files as a winner. Rather it's the users, us, who'll gain the most from the debate in the form of a greater understanding. So please, comment away...