Are web statistics relevant?

August 29, 2006 at 2:59 pm Leave a comment

In my twenty odd years on the internet, statistics for all kinds of services have found their special places. Certainly, from a system’s administration perspective statistical information can be quite valuable. It helps identify the various loads on a network, or individual server, and is extremely handy when troubleshooting various problems. But a whole new market for statistical information was born when the commercial sector finally caught onto the power of HTML and the world wide web. And this market is fraught with error.

Primarily, the big problem with web statistics is that the statistics that are available in themselves are generally technologically relevant. From a business perspective, the information that can be obtained from a web server is not a realistic representation of end user activity on a particular website. From a technical perspective, business relevant information is not well catered for by the various protocols and technologies used for any web application.

Back in the early nineties, the only way to obtain information about visitors coming to a particular website was to trawl through the server logfiles and to attempt to make sense of the information that could be gleaned by the server for each web request. While log files are exceedingly useful to people who look after servers, the information in them is rarely useful from a pure business perspective. For instance, web logs cannot tell you, accurately, how many unique visitors have come to your website. Nor how many visits you have actually had. You can’t really tell anything about your visitors. Not how long they visited for. Nor where they have come from. Nor which individual pages they have viewed.

Sure, you can tell how many requests reached your server, and from which IP address. But hits could be anything from downloading a little graphic on a web page to a web page itself. With some fancy scripting, you can filter out all of the hits that comprise a single web page, and calculate unique page views. And in fact, this became a web marketing standard, a currency, to indicate how much traffic your site was getting. Right until the present day, people still talk about the number of PageViews they get a month. When selling advertising, this is a powerful statistic. But while it remains log centric, the statistic itself is wrong. Its worse than wrong, its meaningless.

Wow! That sorta flies in the face of all those web statistics programs you can buy out there. It also doesn’t make sense. Surely there must be ways that you can actually get this information. I mean, sometimes I go to GMail, or MySpace and it knows who I am?!!! And indeed, there are ways out of logfile hell. But truth be told, they’re not your salvation either.
Slowly,  through the release of articles such as this and this, the concept of Web Analytics shifted from a log centric approach to the use of embedded technologies that would force browsers to give up information about themselves. Heavy use of cookies and other session identifiers began to become commonplace. This helped to make tracking a little easier, because browsers would store visiting information locally and then render it up to the server whenever the server asked for it. Of course, with the proliferation of spyware, and with blunders like this, not everybody uses cookies in the way that you would expect them to. Certainly, I anonymize my cookie to google using this neat plugin. And certainly, there are a host of applications and articles out there to help people clear out cookies and browser hosted junk.

So with this lack of reliability, sites like NY Times will frequently greet you with this page when you try to view news articles. In the hope that they will force you a) to enable cookies to view content; and b) so that they can actually track what you view on their site. And of course, this is simply met with the frequently used BugMeNot tool.

But the whole Web Analytics game is getting worse. With the advent of Web 2.0 (shudder) and the use of technologies such as AJAX, that refresh content items within a page without submitting a new page request, PageViews are rapidly becoming obsolete. In fact, this is nothing new, many flash driven sites are incapable of providing any useful statistical information precisely because of the technology used.

Okay, I’ve ranted about web statistics enough. Why did I even get into it? It seems to me, that all too frequently in my position I am asked to provide information to managers that is actually completely meaningless to them. Truth be told, very few stores count the number of people who walk into them. Certainly, bricks and mortar shops have little hope of tracking “user” activity. From a business perspective, if you’re judging your online business potential on web statistics and not on actual revenue generated from your website, you need to go back to business school.

So, if you’re thinking of asking for some webstats from your IT staff so that you can get an idea of how well or not your website is doing, think carefully about what you are asking for and try to get a basic grip of how the web works before you even come banging on the IT department door!


Entry filed under: Uncategorized.

Hunting for Rare Books in Britain Giving up google

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

Subscribe to the comments via RSS Feed

Recent Posts

RSS New books at Shapero’s

  • An error has occurred; the feed is probably down. Try again later.

RSS New Books at Maggs

  • An error has occurred; the feed is probably down. Try again later.

%d bloggers like this: