Internertnews.com quoted me in their article entitled Benchmark For RSS Client Market Share?, a news story covering Feedburner's RSS Aggregator stats. It's the first time I've been used in the media as a source, so I'm quite chuffed. They didn't contact me, just quoted from my blog - which is fine by me. It's a good write-up and it summarizes some of the caveats involved in trying to measure RSS share. I've been surprised there hasn't been more comment in the blogosphere about Feedburner's stats, but I think that's mainly because people don't know what to make of it. So in this post I'll review a few of the caveats and in future posts I'll address some of the others.
One of the more controversial caveats from the Feedburner post was that some of the Aggregators "ship with one or more of our top 10 feeds as a default" and that this may be skewing the data. Bloglines CEO Mark Fletcher emphasizes this issue in the Internetnews.com piece. He is quoted as saying "...there's a red herring created by default subscriptions (built into desktop software) that can skew results in ways that don't reflect the real user base."
Now, I'm not so sure that default subs is as big an issue as the some of the browser-based aggregator caveats (more on that in a minute). Brent Simmons of NetNewsWire, which was second in Feedburner's list, left this comment on my blog yesterday about the default subs query:
"In the current release version, NetNewsWire 1.0.8, there are 15 default feeds. Of those 15, there are just two FeedBurner feeds -- MacMerc and MacMegasite, same as in 2.0 beta."
He also notes that users can of course unsubscribe from the default feeds at any time.
We don't know how many people do unsubscribe, but given that NetNewsWire is an aggregator exclusively for the Apple Mac OS X then I'd say a lot of users would not unsubscribe from those two Mac feeds. But the real question, perhaps, is how many of Feedburner's big customers are Mac publications? Brent's mentioned two that may be, but how many others are signed up to Feedburner's service? Mac publications are known to have large and devoted readerships, so it's quite possible Feedburner's data is skewed to a degree towards the most popular Mac aggregators - if there are a number of popular Mac feeds in their data.
Mac users big RSS users?
Another thing. If Mac publications are over-represented in Feedburner's data, then this would account for the interesting fact that the second-place getter in Feedburner's stats is an aggregator that is exclusively for people on the Mac OS platform. The vast majority of computer users are Windows users and NetNewsWire is, as far as I know, not an option for them. So it's quite amazing that an aggregator that only a relatively small percentage of people can use, turns out to be the number 2 aggregator for the whole market - and by a healthy margin too! I suspect it's got something to do with Macs being very popular amongst the sort of geek likely to read RSS feeds. Is NetNewsWire the iPod of RSS Aggregators? :-)
Yahoo and the mainstream users
Quick note on MyYahoo, which may be best positioned among the current crop of aggregators to get mainstream users to subscribe to RSS feeds (simply because that's their user base). It's interesting that Firefox Live Bookmarks places third in the Feedburner stats - and my stats for that matter. Firefox users are generally of the geek persuasion. And Yahoo is a fair way back in 9th place, which seems to me another indication that geeks dominate these stats - i.e. Mac users come second, Firefox third! I mean come on, Macs and Firefox are minority products still and it's mostly geeks who use them. I use them :-)
Lest I be seen as picking on the desktop aggregators, I should point out what I think are two pretty big caveats for browser-based aggregators (and let's face it, we're mostly talking about Bloglines!). One issue is that the Feedburner count of Bloglines subscribers doesn't take into account abandoned accounts, another is that a lot of desktop aggregator users do not poll for feeds daily (24 hours is the timespan of Feedburner's study). In the first case, Bloglines stats are probably being overstated. In the second case, desktop stats (like NetNewsWire's) are likely being understated. Those are two pretty big caveats in my opinion.
We've opened up a whole can of worms in this business of analyzing RSS Aggregator market share. But that's a good thing! The RSS world has long needed a way to measure hits and readers. If blogging is to be monetized with advertising and writers getting paid for niche content, then we need ways to measure the stats. How else will advertisers and media companies, and investors for that matter (IPO anyone?) know how to value RSS-based companies and RSS producers?
So that's why I'm so enamoured of Feedburner - I think they've opened up the market for RSS measurement and are leading the way for us all. Interestingly, Feedburner threw out a broad hint that they're about to release a service that will solve some of these stats caveats:
"Clearly, there's a need to dive deeper on stats tracking to start to get a better sense for how widely viewed an item is, how many registered subscribers are actually viewing the content as opposed to just retrieving it, etc. Since we wouldn't mention this unless we were doing something about it, look for a premium offering on this front in the near future."
I've only scratched the surface of the many issues surrounding the RSS aggregator market share data that has been released by Feedburner. But that's OK, because it means I get to keep digging into and analyzing the data! Now, Feedburner: about those stats minus the top 10 feeds... how about giving those to us? ;-)