RSS Market Share

There has been much discussion and lively debate in the blogosphere and other spheres of note lately regarding RSS traffic and aggregator market share. Since our job is to help publishers understand how their feed is being consumed in as accurate a way possible, we've decided to make our own understanding of traffic patterns and market share public. As with any micro-report like this, there will be much gnashing of teeth and waving of hands, and in the end, this is just as likely to confuse as many people as it helps. Transparency can only assist the discussion, however, so here we go.

First, a lot of context. There are a number of aggregators that currently consume feeds on behalf of multiple subscribers. Some of these (eg, Bloglines, My.Yahoo, NewsGator Online) relate the number of subscribers on behalf of whom they're polling when they make http requests, but others do not. Therefore, straight away, we can tell you that the market share numbers are skewed to the extent that the currently non-reporting proxies have sizeable numbers of subscribers.

Next, an aggregator's ping/hit frequency has zero bearing on its market share. FeedDemon, for example, is an extremely efficient aggregator and makes far fewer polling attempts with no loss of timely data updates. Our market share data below is based on aggregate circulation as of January 6, 2005 across the top 800 feeds we manage. FeedBurner measures RSS circulation by creating unique combinations of user-agent and IP addresses within a single 24 hour period and then factoring for anomalies (example anomaly: the Danger hiptop for T-mobile user-agent comes through a single ip proxy on behalf of all its users, so that's not a circulation of 1). So, if an Opera user-agent at ip address X polls a feed every minute on the minute all day (to use a realistic example), we only count that as a circulation of 1. This has obvious limitations, but it allows us to create a reasonably accurate view over time for a feed's circulation.

It should be pointed out that web-based aggregators may have an overstated market share to some unknown degree, because if there are 100 web aggregator subscribers to a feed, we are counting all of them every day when the web aggregator checks in and tells us the feed has 100 subscribers, regardless of how many people actually read the feed that day. 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.

Ok, we could go on and on with context, but instead we'll just make this topic a regular post here, and each time we'll explore some other piece of the puzzle. Now, some key summary points and then some stats.

Key points:

• RSS Client market is not yet consolidating, it's expanding. There were 409 different clients polling the top 800 FeedBurner feeds in September and now there are 719 different clients. FeedBurner actively catalogs the behavior and specifications for hundreds of these user-agents.
  
• Mobile RSS readers are already growing in popularity. While none yet crack our top 20 list, there are thousands of users of the FeedBurner Mobile Feed Reader alone.
  
• Of the feeds that we currently manage, in aggregate, RSS circulation is growing by about 1% every weekday. We are also adding to our total number of feeds under management at the rate of about 1% every weekday. If the widely discussed Pew research estimates are correct, we currently reach about 5% of the total RSS subscribers in the US, although we currently manage just over 1% of the world’s active feeds.
  
• We currently manage a significant percentage of the world's active podcasts with our SmartCast service. The podcasting phenomenon is accelerating. Enclosures are an important trend. More on all this in a separate post.
  
• We have been producing this list internally for quite a while, and it can and does change dramatically in a short period of time, both because of the changes in the sorts of feeds we manage and the changing trends in client popularity.
  
• This list is heavily skewed toward aggregators used on blog feeds, since most of our feeds are from blogs. This list might read quite differently for more traditional media feeds such as Reuters, NYT, CNET, etc. On a similar theme, individual publishers will notice that the overall market share may be wildly different from their own feed's market share. Simply removing our top 10 feeds from this data results in a wildly different market share list, possibly because of clients that ship with one or more of our top 10 feeds as a default. All of this pointing to the caution not to read too much into this single data point. We could make qualifications about everything on the list. Your mileage may vary, caveat emptor, mea culpa, c'est la vie.
  

Top 20 RSS clients across FeedBurner most highly subscribed 800 feeds as of January 6, 2005

Aggregator Name (Market Share Percentage)
1. Bloglines (32.86%)
2. NetNewsWire (16.95%)
3. Firefox Live Bookmarks (7.78%)
4. Pluck (7.20%)
5. NewsGator Online(4.45%)
6. (not identified) (4.07%)*
7. FeedDemon (3.83%)
8. SharpReader (3.27%)
9. My Yahoo (2.58%)
10. iPodder (2.42%)
11. NewsGator (2.23%)
12. Thunderbird (2.13%)
13. RSS Bandit (1.12%)
14. NewsFire (1.05%)
15. iPodderX (1.02%)
16. Sage (0.71%)
17. FeedReader (0.67%)
18. RssReader (0.54%)
19. LiveJournal (0.46%)
20. Opera RSS Reader (0.45%)

* (not identified) represents an amalgam of clients that request feeds with a blank user-agent field. This total likely represents tens or hundreds of different clients and is not one particuarly large stealth project.

Same list in chart form for clients with > 1% share