Comments for The State of the Web-based RSS Reader Market - Feedburner, Pheedo Release Stats

The State of the Web-based RSS Reader Market - Feedburner, Pheedo Release Stats

2007-02-22T21:40:17Z

Feedburner has released an interesting new report on web-based RSS Readers, prompted by the recent introduction of Google Reader into its stats (incidentally, for some reason R/WW only increased by around 20% after Google Reader was added to Feedburner; whereas most other tech sites increased by 40+%).

RSS Analytics plagued by unreliable stats

Feedburner is putting a lot of effort into enhancing the way it measures RSS feeds. I was talking to someone the other day about how RSS analytics is still very much a nascent industry - i.e. it's even more difficult to get reliable feed reader stats than it is to get reliable webpage stats (and I've written before about how easy it is to manipulate both). Feedburner itself points out one of my particular bugbears at the moment - how being a default feed in an RSS Reader like Netvibes or Pageflakes artificially increases your RSS subscriber number (in some cases by a large amount). R/WW has benefited from this behavior too, as we are a default feed in the bundles that Rojo provides. But alas, we're not a default feed on the other popular RSS Readers and startpages :-( Anyway, all of these things mean that a blog's RSS subscriber number should be taken with more than a few grains of salt.

Feedburner focuses on Audience Engagement

Despite all these issues with RSS analytics, Feedburner is leading the way in feed management for consumers and in this new report they discuss some new measurements - focused on how people are reading feeds and interacting (i.e. clicking or 'viewing') with them. They call this "Audience engagement" and it seems to be a priority now for Feedburner, in order to counter the issues with straight subscriber counts (e.g. the default feed issue discussed above). For the record, I think this new focus on audience engagement makes perfect sense for Feedburner - as it will help move RSS analytics forward and remove some of the stigma attached to it, due to the current unreliability over RSS subscriber counts.

]]> Feedburner has an aggregate db of 604,533 feeds and from that, they've come up with the following chart of audience engagement by 'views', for web-based RSS Readers:

It shows that Google Reader is now the number 1 web-based RSS Reader with 59%, well ahead of former king Bloglines (which has been cruising along with no major re-designs, living off its reputation, for too long now). Netvibes has shown amazing growth over the past year and is now in an enviable position amongst the start pages. Newsgator is probably more of a niche enterprise and high end consumer play these days, which may suit it just fine. In reality, it's a battle between Google and Bloglines (owned by Ask.com) for web-based RSS Readership.

Note that MyYahoo, one of the top 3 RSS Readers overall in Feedburner, does not figure in 'views', as it only displays content summaries - so users need to click through to the site to view the full content. This is reflected in Feedburner's graph of audience engagement by clicks:

A similar thing could be said about Netvibes and the other start pages, where users generally click through to view content. Interestingly, both of these charts suggest that Netvibes is 3 times larger than live.com, Microsoft's start page. Although I'm sure even Microsoft would admit that at this stage, Netvibes is a much more sophisticated product. Microsoft (and Google) both probably figure they have plenty of time to catch up, as start pages are still a niche early adopter thing. In fact, there's a very good chance Netvibes or Pageflakes will be acquired by one of the big 3 (Google, MS, Yahoo) by end of this year. My money's on Yahoo acquiring Netvibes.

Pheedo's Stats

Pheedo has also come out with some stats for web-based RSS Readers. Their chart for market share indicates that Newsgator Online leads in subscriber numbers (in Pheedo):

Note however that Pheedo's stats for Spring 2006 (PDF) showed that Bloglines was leading with 30% share then, against just 10% for Newsgator:

So what has changed since Spring 2006, to make Newsgator Online the leader amongst Pheedo publishers? I think that needs more explanation, if anyone from Pheedo is reading this.

Summary

It's clear that Google, Yahoo, Ask.com (Bloglines) are leading the way with web-based RSS Readers. Google and Yahoo have different approaches for now and Microsoft isn't bothered with a general web-based Reader - they will focus on email (Outlook), the browser (IE) and start page (live.com). Netvibes is showing impressive numbers and so must now be a prime acquisition target for one of the big 3 - and my guess is that Yahoo needs a strong 'start page' type technology moreso than the other two, who are doing it themselves (live.com and Google Personalized Homepage).

Disclaimer: I am an advisor to Nooked, which is an RSS marketing company.

Comment from Bill Flitter on 2007-02-22

2007-02-23T00:52:33Z

Richard,

First, thanks for including our data in your post.

"So what has changed since Spring 2006, to make Newsgator Online the leader amongst Pheedo publishers?"

RSS analytics measurement, as you know, is a tricky business. We are constantly evolving how we view the data coming to us from the RSS aggregators, what we count, and when we count it. And based on what Feedburner is reporting, they too are. The reason for the change from 2006 to 2007 is exactly that. We updated our algorithms and strategy how we count and report on the data.

Cheers,
Bill Flitter
Pheedo

Comment from Stoicho M on 2007-02-23

2007-02-23T22:38:22Z

As a everyday feed reader user from 2 years, I have used Bloglines -> Live.com (early stages) -> Rojo -> Google Reader (one week test) -> Rojo -> Bloglines.

I have never heard of NewsGatorOnline, although I know about Newsgator. This is nonsense to me: "We updated our algorithms and strategy how we count and report on the data."

Did you updated the algorithm or did you updated your business strategy on how you count? Or both?

Comment from Richard MacManus on 2007-02-23

2007-02-23T22:58:42Z

Thanks for your comment Bill, but I'm with Stoicho on this -- it doesn't really make a lot of sense. What specifically changed in your algorithms or strategy that saw NewsgatorOnline leap from 10% to 27% (to make it #1 for Pheedo publishers) in the space of a few months? Feedburner data suggests that NewsgatorOnline is very much back in the pack in terms of its market share, so there must be a reason why it's number 1 in Pheedo?? Can you illuminate more?

Comment from Bill Flitter on 2007-02-24

2007-02-25T00:38:14Z

Stoicho - NewsGatorOnline is one of NewsGator's aggregators. They have a family of readers that include NewsGator Enterprise, NetNewsWire and FeedDemom. We did not report on Netnewswire and FeedDemon for this comparison. Secondly, we have no incentive in how the numbers turn out. We report
this data for the benefit of the community.

Richard, Pheedo's market share data and Feedburner's engagement data is not comparing Apples to Apples. We are looking at circulation data (explained below) and reach (see data at http://www.pheedo.info/archives/000401.html) to give an indication of market share. Feedburner is looking at views and click data to determine market share. Just two different ways to approach the situation. SimpleFeed uses unique URLs to aid them in counting subscribers.

Market Share Calculation/Subscriber Count Data:
SimpleFeed = Unique URLs
Feedburner = Views and Clicks
Pheedo = Circulation counts based on subscriber data and algorithms
Nooked = Richard or Fergus, maybe you can answer this?

Without getting into too much detail or proprietary information, below is an example scenario to shed some light on the complex problem.

Hour 1 Aggregator A reports Feed 1 has 5 readers
Hour 2 Aggregator A reports Feed 1 has 4 readers
Hour 1 Aggregator B reports Feed 1 has 3 readers
Hour 2 Aggregator B reports Feed 1 has 2 readers

How many readers does Feed 1 have at the end of the day 5+4+3+2 = 14? Or do you average the numbers for A & B to determine an average subscriber count? In late 2006 we put together a proprietary technology to get us closer to a unique count. We set out to determine, of those 14 subscribers reported by Aggregator A, how many of those were uniques. As I said before, we are constantly improving how we approach the situation. Our incentive is to deliver the best possible information to our marketing partners. Subscribers are ONE stat of interest to our partners along with impressions, clicks, where the feed is syndicated, and how it is being used. And we believe, to look at the growth in the Google Reader, using subscriber count was the best comparison. As Google said, it was a major user request.

Also, keep in mind, we report circulation vs. subscriber counts. Using a newspaper analogy:

Subscribers: Readers who have paid to receive a paper.
Circulation: Readers who have paid to receive a paper plus all the papers dropped off at the corner news stand.

An important distinction. We use circulation because, some readers tell us number of subscribers and other's don't so we have to rely on some basic web technology to derive that number - much the same way website analytics providers derive uniques to a website. We've thrown a lot of technology and methodology at the situation to derive an accurate count.

You argue, user engagement is a better measurement. I think it is one part of the calculation definitely but not doesn't give the full picture. For example, consider full-feeds vs. summary feeds. CTR on full-feeds is lower but it doesn't mean a user is any less engaged. They get all the information they need at the aggregator and don't need to click through to the site. Rick from Feedburner has a good explanation in the comments at Blogspotting (http://www.businessweek.com/the_thread/blogspotting/archives/2007/02/feedburner_myya.html)

The question is what reader has the biggest market share? It would depend upon what you believe determines market share (Clicks or Views or Subscribers or Unique Feed URLs).

Thanks,
Bill Flitter
Pheedo

Comment from Richard MacManus on 2007-02-24

2007-02-25T05:20:26Z

Thanks for the info Bill, it helps understand how the different companies measure RSS. But, with respect, it still doesn't explain why NewsgatorOnline jumped from 10% to 27% in Pheedo's stats ;-) There's no ulterior motive, I am just curious if there's any reason you can identify as to why they're #1 with Pheedo publishers.

Comment from WoW!ter on 2007-02-25

2007-02-25T15:09:28Z

I think a real problem with these stats is that cumulative usage is measured, and not active usage.
I think that all my Bloglines readers have moved to Netvibes, so I doubled my Feedburner stats, since both are still reported.
Another intersting addition to this post would have been the hitwise blogpost on Bloglines vs other readers (Google reader) about mid january.
I just did that at my blog.

Comment from Fergus Burns on 2007-02-26

2007-02-26T13:35:21Z

I posted previously on this topic

http://blog.nooked.com/archives/2006/08/measuring_rss_c.htm

from my perspective
- publishers should analyze feeds on the basis of "reach"
- Pheedo originally created the idea of "circulation"/"reach".
- Feedburner moved from "circulation" to "subscribers"
- the "black art" calculations for this "subscriber" count is misleading

Fred Wilson's recent comment backs up my view on this
[see http://avc.blogs.com/a_vc/2007/02/feed_stats_what.html ]

"So what to do? It's really simple. Focus on "reach". FeedBurner calculates a reach number every day for every feed it manages......I suspect the same is true of Techcrunch's 279,000 readers as well. If their feed subscribers are like mine, Techcrunch's daily reach would be about 50,000. That's still a huge number of people reading a feed every day."

From our perspective, we offered our marketing customers the option of providing a "broadcast" feed or a "personalized" feed.

Almost all our customers are choosing "personalized feeds" - as this fits in with their current marketing metrics model, and they can track the conversions at a granular level.

From a technical perspective, we've reduced some of the redundancy for the feed readers [see http://blog.nooked.com/archives/2006/04/personalized_rs.htm ]
but vendors need to get on board.

See an old Scott Gatz post on the subject - http://www.scottgatz.com/blog/2005/12/13/last-nights-rss-marketing-industry-roundtable/

Hope this helps in understanding a bit more @ feed measurement :-)

Best Regards
Fergus