I was doing some thinking about measuring blog usage today—not how many blogs there are, but how far blog content reaches. Such measurement isn’t a priority for most blog publishers, but what about traditional media companies that have to decide whether to make the RSS plunge as a business investment? So I came up with a few observations:

1. Reach is a traditional media measurement that calculates how much of the potential viewership (I would use the word “audience,” but we all know that’s a screwed up metaphor for online activity) can see a particular piece of content. This is a hard measure to get, since the content is exposed not only on one’s website but in an XML file that can be exposed in lots of different readers, and when the headline can be posted on lots of different sites. Other traditional media metrics based on exposure, including unique users and cost per impression, also go right out the window.
2. Likewise, coming to a decent clickthrough measurement is difficult, since clickthrough is defined as clicks per people that viewed the link (see above).

So what does that leave us with? What about treating RSS like newsletters? Subscriber count is hard to gather since there’s no “formal” subscriber process to get an RSS file. Likewise download count for the RSS file: while the latter is feasible, platforms like Manila don’t render a static XML file that can be tracked in a traditional web hit log, and counters like SiteMeter only track files that can embed their counting code (which leaves out RSS). And it’s hardly a meaningful or reliable measure of exposure without unique users, or knowing whether the downloaded file actually contains new content.

So what’s a media company to do? Other than take it on faith, I mean. Maybe starting with Technorati? Or Google’s PageRank?

Too many questions, not enough good answers.

And then there’s this…

A reader emailed after my post last week about measuring RSS to ask “Why not slip a 1 pixel ‘webbug’ into the RSS feed?” Good question.

Advantages of web bug graphics:

• Unobtrusive in the RSS reader’s pane.
• A direct hit to your server, allows you to play games like feed views and unique users. Note I didn’t say page views or content views; more on that in a second.

That’s about it, really.

Disadvantages? Plentiful:

• No referrer, no specific content tracking. RSS readers generally send “no referrer” per the HTTP standard rather than try to make up a referring URL (though some, like Radio Userland, refer back to a host page for their services). So you can’t track which content piece the reader was coming from.
• Doesn’t always get forwarded. RSS items generally contain minimal markup, so an extra image tag inserted is sure to be noticed and removed by most bloggers. Why is this important? We care about the total reach that our content gets, on other peoples’ sites as well as our own. At minimum you won’t be able to count any references that your content gets on tech-savvy bloggers’ sites.

Plus, of course, any tracking system that relies on client side code can be exposed—and risks your readers’ alienation. And, as we’ve discussed before, alienating your blogging readers can be a sure way to invite shunning—and shrink your reach, but good. Kind of the opposite of what you’re trying to do in the first place.

And, by the way, this goes double for any more complicated embedded Javascripts or other solutions.

But what about embedding meaningful images, each with a unique name (perhaps associated with your article’s GUID), in each article? Kinda suggests that the photobloggers are the people most likely to get real tracking of how their content is read.

Of course, they’re also the most likely to get it stolen, renamed, and rehosted on someone else’s site.

This is starting to feel like the three laws of thermodynamics, which I propose we recast as the three laws of measuring RSS:

1. You can’t win.
2. You can’t break even.
3. But you don’t want to get out of the game. Not if weblogs are worth one one-hundredth of the hype that they’ve received. (And I think they’re underhyped. Weblogs at Harvard? You ain’t seen nothin’ yet.)

A follow up to my thoughts about the acceptability of inserting arbitrary markup into RSS feeds to measure usage. Some RSS “readers” just display headlines (such as the Radio and Manila RSS Box), meaning that the tracking code would have to be in the title element of the RSS to measure exposure successfully.

But Mark Pilgrim’s experiment last week has awakened the authoring community to the danger of arbitrary markup in RSS, and it appears the community has quickly decided that titles aren’t for markup.

Why did we go down this thought road in the first place? So we could track page views of RSS content. Why? To get clickthroughs (total clicks divided by total pageviews, for each RSS exposure). But we can’t get clickthroughs that way.

How about this: the most effective way to measure RSS usage is to put a tracking URL in your RSS feed, one that’s distinct from the one you expose through your navigation. This should be trivial with a good content management system (which all blogging engines are), and you needn’t even make the tracking URL hop through a redirect. On the landing page you can count unique users and all that fun stuff, and if the RSS link has been posted to other pages for discussion and people click through there you’ll be able to track the spread by watching referrers.

It can’t be that bad a system—after all CNet uses it. All links in their feeds include the parameters “part=rss&tag=feed,” and some even are directed to a special host, rss.com.com (oddly, the home page for this redirects to download.com).

In short, we don’t get reach and we don’t get clickthrough rates. But we get something that can empirically measure the effectiveness of RSS against other content promotion technologies.