Lies, damn lies, and web statistics
Posted April 27, 2007 at 12:47 pmAnother Beth blogs:
Web Analytics packages are sold as if it’s an automatic coffee maker. In fact, it is more like buying a coffee plantation. You can still get your cup of coffee (eventually), but your [sic] going to have to stick your hands in a lot more manure than you ever knew.
How funny! (And how true.)
Ten years ago I wrote a paper for an STC conference titled “Lies, Damned Lies, and Web Statistics” that talked about doing web metrics in the days well before there were “real” web analytic packages.
And the funny quote then was:
Interpreting web statistics has been described as trying to nail Jell-o to the wall.
My approach back in the day was to write Perl scripts to churn log files. How passé now! But many of the issues (such as what you actually measure versus what is interpreted) are still very valid, and newer technologies, like AJAX are only muddying the metrics waters.
April 28th, 2007 at 6:20 pm
On a related note, did you the very interesting IAB open letter to Comscore and Nielsen ?
In a nutshell, internet panels are still smack:
http://www.iab.net/news/pr_2007_04_20.asp
“The glory of interactive media is they make it easy to assemble, count, and assess the marketing value of these and myriad other niche populations – and aggregate these niche populations into effective and efficient media plans.
To persist in using panels that undercount or ignore the diverse populations that are the future of consumer marketing is to deny marketers the insights they need to build their businesses.
And it certainly appears to us as if they are being undercounted or disregarded, for our members’ server logs continue to diverge starkly from your companies’ sample-based assessments, by 2x to 3x magnitudes in some cases – far beyond any legitimate margin of sampling error.”