Traffic Stats are Dead
So says Steve Rubel. And a bunch of others.
To be dreadfully honest, my knowledge of the intricacies of web analytics is infantile at best, but what I do know without a doubt is that they’re often misreported. The old business adage goes, turnover is vanity, profit is sanity. As Steve says, “The advertising economy is built on reach. It’s time for it to change to depth.”
So what’s wrong with page views (the incumbent metric-of-choice)?
A page view is each time a visitor views a webpage on your site, irrespective of how many hits are generated. Pages are comprised of files. Every image in a page is a separate file. When a visitor looks at a page (i.e. a page view), they may see numerous images, graphics, pictures etc. and generate multiple hits.
Evan Williams explains it really well (for non-techies):
Remember when web site traffic was talked about in terms of “hits”? You’d read about how many millions of hits Netscape got per month and other sites bragged about getting 30,000 hits a day. Eventually, we moved away from the term hit because everyone realized it was pretty meaningless. You see, a hit was often counted (depending on who was counting them) not just for a page load, but for every element (e.g., graphic) included on the page, as well. One visit of this page, for example, would be worth about 40 hits (if the browser had images turned on). But a site that was less graphical and had equal usage would register half the hits.
Pageviews replaced hits as the primary traffic metric not just because they’re more meaningful, but because it also determined how many ads could be served. Ads were sold primarily on a CPM basis, so multiply your CPM by every 1,000 pageviews you got, and that’s your dot-com revenue.
Reach (number of unique visitors) is also important, of course. comScore/Media Metrix uses uniques as its primary metric, because mainstream advertisers want to reach a lot of people, not just the same people over and over.
Page views don’t count for RSS subscriptions / readership, are rendered ineffective for sites built with Ajax and in a widgetized Web where bits and pieces of numerous sites are splattered over bits and pieces of other sites – a gloriously messy social humdrum of activity. But is any of it valuable?
So the social Web, businesspeople or not, are calling for more meaningful website metrics. In an age where social software has allowed us as individuals to create highly niche-oriented sites clusters of similar social sites could be extremely valuable to marketers if they had metrics for measuring social value. Culture? Trust? Respect?
Any ideas how we could more accurately measure the value of websites? And let’s say for the purposes of this discussion, although many of us don’t advertise on our sites, that we were convincing an advertiser…
I think a sites worth can be measured in a few areas, depending on the type of site.
1. Blog
The amount of quality posts matched with users comments on those posts. Eg. Techcrunch
2. Social Sites
The user base and the frequent activity of those users. I mean its useless having 100 000 users registered on your site if only 100 of them actually contribute. Eg. Flickr & digg
3. Search Engines
Well this is an easy one, on the amount of queries handled and the variation of those queries, unless it is a niche search engine. Eg. ThatShoe by Incubeta.
P.S. Mike please make this comment box bigger. :)
Thanks for the comments Tyler – comments box will magically expand in size in the next few days :)
I think Google has the right idea (once again). Their “Site overlay” feature on Google Analytics allows you to measure mouse clicks mapped to certain areas of a page. Presumably this is through some javascript/ajax wizardy. It’s bound to be more effective than page views in the AJAX/dynamic page age. It’s still far from perfect, but if some standard can be built around measuring this metric I think it’s the most reliable current measurement of site activity.
RSS is a fair bit trickier, but the final solution (if there is one) could require a combination of RSS subscription hits with the above “page click” count. This still leaves the problem of duplicate visits which need to be filtered based on, I don’t know, maybe sessions or something (thinking of users behind a proxy here). A session id on the RSS feed could help.
Hmmm, can’t solve the net’s problems yet – lemme grab another beer and think about it.
I agree with all the analytical and measurement problems associated with current metrics — I used to crunch nunbers for a living — but I am less concerned about these issues since I don’t sell advertising.
Plus, having been tracking Google Analytics for over a year, I’ve noticed a pretty high correlation between page views and unique visits (I don’t have multipage documents).
I’m basically interested in connecting with people via my blog. So I care about relative page views (which is why I publish a “top ten” list every week) plus I care about how people navigate (which has led me to modify the layout and contents of my inital page and the links available from all pages).
So I guess when I wear my statistics hat I am ready to criticize current metrics (http://www.ddmcd.com/rattle.html) but, in the scheme of things, I don’t really care that much.
As usual, Scrivs has a good take on this issue.