One of the responsibilities that my new team has is to extend the reach of information about the Microsoft developer tools and platforms through their blogging and other web activities. We've blogs sitting on the long-standing MSDN and TechNet services, and a site of our own at http://www.ubelly.com/.
Conversations last week turned to the challenges of calculating readership of these services given the apps and RSS world in which we live today.
Not that calculating readership has ever been easy. In the pre-Internet world of printed newspapers, The Audited Bureau of Circulations (ABC) gives a certified number for the distribution of a title. But distribution isn't the same as readership (how many people might look at a particular copy; how many copies are never looked at at all?), and these slightly flawed metrics were taken into the Web 1.0 world a decade or so ago when web advertising was based on the idea of CPM (Cost per Mille – how much to get an advert in front of a thousand viewers).
These days, however, working out what readership of a particular site or article might be is increasingly complex (if impossible). To illustrate, let's look at the stats for this blog in March…
The stats that are generated by Blogware, which is the product this blog is hosted on tells me that in March 2011, there were 5,438 visits to this site, with over 11,800 page views. Impressive stuff, perhaps? But according to Google Analytics, which does a whole load of filtering to give information about “real people” viewing my site in a browser, the numbers drop to 561 visits with 873 page views… (and it also tells me that there were 308 unique visitors in that month too).
Why the discrepency? Well, because Google Analytics filters out a stack of machine requests, and lo and behold, the top five browser types according to Blogware in March were search engine bots, making up about 50% of the page requests in the month. However, at number six (and number eight) in the browser list are the Feedburner and Google Reader bots, and that is where calculating readership starts to get really challenging these days. Those RSS reader services are presumably getting the content so that someone can read the article either online in an aggregator, or on a PC or mobile app, but get discounted by Google Analytics because they are machine derived…
Now the other thing that needs to be kept in mind is that readership and somebody actually reading something are also different things, and that in fact no amount of web-traffic-based analytics actually tells you that someone is reading (let alone comprehending) something that you have written. For that, you need to get them to actually do something. That, of course, is one of the areas where the web has revolutionised the world in the past 15 years, by turning advertising from a business of selling space to a business of selling leads. That readership therefore has become a less sought-after stat probably is going to make our lives that bit harder…
mmitII 