Some of the more commercially-minded members of the blogosphere got together at the Blog Business Summit in Seattle this week, and although I wasn't there, I've been following along from afar. One of the more interesting (to me) discussions was about measuring your audience and your reach; as vital a topic for Boeing as it is for Boing Boing, or any other corporate bloggers or upstart media concerns trying to make a go of microcontent as a living.
The summary of this chat was really that there is no silver bullet. There isn’t anything in blog/RSS analytics that tells you everything you need to know and tells it accurately.
An audience member added another point, “I don’t care about how many page views or visitors I really get. I care about getting the right visitors, the influential visitors, or the potential customer visitors. How can I tell who’s who?”
Having done some time in the online advertising technology and analytics world (I did a two-year bid at 24/7 Real Media before I got early parole for good behavior), I can tell you these concerns extend far beyond the blogging world. Every media company is struggling with the transformation from unnacountable advertising to very accountable ad models.
In the past, advertisers had to make due with not knowing how their marketing really peformed, trusting instead in audited audience numbers and qualitative surveys to tell them who a given media outlet was reaching. You want 18-34 year olds with a lot of discretionary income? Buy 30 seconds on Friends. Men aged 18-34 looking to buy a car? That would be Monday Night Football. Fat-cat investor types? The Wall Street Journal.
None of these media properties, of course, could prove that any of the advertising they carried actually did what it was supposed to do—sell stuff—but the advertiser took some comfort in the fact that there was nothing better they could do (short of direct mail, but that's got the stank of Valu-Pak all over it). The rise of very accountable online advertising, along with the cost-per-click model that GoTo/Overture/Yahoo! poineered (and Google perfected) is destroying that blissful ignorance.
This puts a lot of pressure on publishers to come up with the goods on their audience: how many? what do they look at? where do they come from? who are they? when do they show up? Not easy questions to answer.
Server logs are totally inadequate to the task, so no help there. JavaScript-based web analytics tools, whether they're the freebies from Google (based on the former Urchin product), or the heavyweight services from companies like Omniture (what GM's using to track you when you read their blogs or visit a GM site, incidentally) are far better bets. Even so, what can they tell you about who an audience member really is? it's an impossible dream. However, an emerging class of analytics-driven targeting services out there (like TACODA and Revenue Science) aim to do the next-best thing: target based on what a unique visitor does, both on and away from a publisher's site. In this way, someone who spends their day surfing videogame oriented sites would be shown ads for related products.
There are shortcomings, of course. To begin with, many people just don't want to be tracked this closely (even if the tradeoff is less irrelevant advertising). They surf with third-party cookies turned off, and with their browsers or firewalls set to block content from certain domains (I'm looking at you, doubleclick.com). And, as a small-scale publisher of a blog, the network that sells ad space on your site may, or may not, employ this kind of targeting. On the other hand, the advertiser isn't necessarily buying an individual property, they're buying a targeted segment (like "Web 2.0 buffs"); it's up to the ad network selling your inventory to classify your site into the right package.
So, ultimately, analytics are interesting, but not necessarily a primary concern to the blog author. A better use of our time is ensuring what we write is worth reading.
Uh-oh.
Tags: advertising, blogs, media, analytics