More thoughts about measuring Web 2.0
Clint was able to comment on my last post late last evening but I tried to respond in comments but apparently something is wrong at Blogger so I’ll respond in a post and offer that if you’re trying to comment on my “measure Web 2.0” post you can email me directly and I’ll publish the comment for you somehow.
Clint’s comment “Fie! The page view is NOT canonical in a web 2.0 world.” is spot-on correct and that’s a big part of the problem, I agree. But, and perhaps unfortunately, page view is the simplest and most significant form possible without the loss of generality offered by the vast majority of web analytics applications. So sure, in the AJAX and RIA models most of the interesting stuff happens below the level of a traditionally defined “page view” but if you don’t have any good way to measure that does it matter?
My proposal for measuring Web 2.0 assumes an end goal of integrating all site data into your current web analytics platform. For obvious reasons I’m not likely to advocate for additional silos of data, all hell bent on describing the exact same visitors in a completely different way.
So Fie! Back to you Mr. Ivy. I actually believe that the most basic event type in Web 2.0 is, and don’t laugh here, the “hit”. The same metric that I described in Analytics Demystified under the header of “A Small Group of Mostly Useless Terms” and that I recently told a packed room at Search Engine Strategies was an acronym for “How Idiots Track Success” is potentially returning to grace as critical to the measurement of this emerging application development model.
You’re laughing, right?
Seriously, the definition on page 46 in chapter four cites WebTrends saying “[A hit is] an action on a Web site such as when a user views a page or downloads a file.” Back in the day the fine folks at WebTrends almost surely didn’t envision measuring remotely located web-based applications like Google Maps but by capturing the idea that a hit is an action you can hopefully see where I’m going with this.
All we’re really trying to measure from Web 2.0 is actions which ideally tell us something about our visitor’s engagement. From that we can build a handful of swell key performance indicators like “Average Actions per Widget” and “Average Time Spent per Widget” and “Percent Widget Users” and so on.
Speaking of widgets, my boss serendipitously sent me a link from Fred Wilson pointing to information about how the Zevents widget supports web analytics. The guys at Zevents are absolutely thinking about it the right way but I think they need to take it a step further and make the information available via the strategy I described in my last post. They have a handful of good events described (“event views”, “venue views”, “searches”, etc.) but again, I think that the great utility of this type of reporting will come when I can integrate the data into my existing solution (whatever that may be)
So kudos to Zevents, fie to Clint and keep the comments (and email) coming!