The end of the RSS experiment

Last week we reported on our site statistics after going to a full RSS feed. The results were disappointing; our numbers went down. We said we'd continue the experiment for another week to see if the trend was reversed once more people heard about the option of viewing all CogDaily content in RSS feeds.

Now, after another week of full RSS feeds, we have more results to report. These results confirm what we found last week: while RSS subscriptions are up, page views are down:


The dotted line represents the date we switched to full RSS feeds, and as you can see, the trend continues. Last week was a bit of a light posting week due to a computer glitch losing one of our Research posts, but even accounting for that the numbers are still lower. Here's the graph of several different stats using z-scores to make them comparable:


As you can see, the overall trend has been downward. So over the short term, there's a fair bit of evidence that full RSS feeds reduce activity on a blog.

This little experiment doesn't really address the long-term issues, and we are still very actively considering the possibility of instituting permanent full RSS feeds for Cognitive Daily supported by small ads in the feeds themselves. However, a number of issues would need to be worked both by Greta and me and by, and so for the time being we're going to switch back to the old, partial feeds. We think partial feeds still offer a number of advantages for readers -- and, of course, you can always come to Cognitive Daily to get the full post.

More like this

For the past week, we've been conducting a little experiment with Cognitive Daily. In the past, we've had several readers complain that we don't include the full post in the RSS feed for CogDaily, so last week we published every post in its entirety on RSS (if you don't know what RSS is, I explain…
Last week, the number of subscribers to this blog's RSS feed passed the 2,000 mark, after teetering just below that number for a couple of months. It's very gratifying to know that so many people enjoy my writing, and although this increase in subscriptions is tiny compared to the total number of…
One of the most hated practices on the Internet is the breaking of articles into pages. Jason Kottke swearingly rants against it here, and Mike Davidson denounces the practice here. I don't much like the practice either, especially when a short, pointless article is broken into four or more pages (…
Very few relationships in this world are monotonic. Not the price of stocks, not the traffic on this blog, and not global climate trends. Maybe if more people understood this, we'd have less nonsense about climate change clogging the media. By monotonic, I mean, if you plot a trend on a standard x-…

There is definitely a downward trend in pageviews, but that is to be expected. If a reader can view the whole post in a RSS reader, then there is no need to open CD in a web browser. I don't see a downward trend in either links or increase in RSS subscribers. The only downward trend that seems noteworthy is in the comments, and then only in the last week where you admit to it being a light posting week. I would posit that the number of comments from May 14-20 is highly related to the number of posts, and not related to the fact that you changed to full RSS feeds.

I know almost nothing about statistical analysis.

Before the dotted line the rate of change on the page views was very steep and negative. After the dotted line that negative trend is halted. One interpretation would be that the change halted a drop-off in page views.

However you look at it, page views were going down prior to the RSS format change, so attributing the negative trend completely to the change does not make sense.

Again, I know nothing about statistics, but I think that for these numbers to make any sense you would have to do some analysis of the trends for many weeks before the format change to determine if the trend in this sample is significant. You would also need to repeat the experiment at a later date to determine corriliation.

By codesuidae (not verified) on 21 May 2007 #permalink

Well, since page views are what we're after, until we can develop a system for putting ads in the RSS, that's really the deal breaker.

Adverts in the feed is so wrong-headed: the goal is more pageviews thus improved content is the obvious choice. Partial feeds demand a good first paragraph that grabs attention, concisely describes the full article and invites that all-important click.

(When I read a good article in a full rss feed, I click through to the article to register my appreciation.)

By Matt Platte (not verified) on 21 May 2007 #permalink

I've only been reading this blog for a week, through my Google Reader. The best part of it for myself, is no ads. You'll be losing a reader if you stick ads in the feeds. Otherwise, keep up the good work.

This is the evolution of the internet, page views go down, comments go down but subscriptions go up. People get their news faster, more conveniently and no longer have to type in your URL or even click on a bookmark. It's the best way to read news.

Maybe you should look into other revenue ideas, so as not to alienate your new and returning users. The information you have provided in the graphs is definitely interesting and raises various questions on how to keep your site "alive".

I wish you all the best of luck in this endeavor.

I read the blogs about 50% for the post and 50% for the comments. For that reason, I don't use the RSS interface although I know how to and have used it in the past, but ultimately found it unsatisfying.

I think that comment participation by the blogmaster is fairly important to encourage participation; unfortunately, many blogmasters have taken to using "low hanging fruit controversy" to bump the comment participation. Stooping to sensationalism is unnecessary in many cases, but I guess it depends on the target audience.

I think full RSS feeds with advertisements kind of negates the pupose of RSS feeds in the first place: A summary that affords one to choose whether or not to read the full article. At least that's how I use it: quick access to the headlines, without the hassle of all the ad-related scripts that slow down the reading experience.

By Jan-Maarten (not verified) on 21 May 2007 #permalink

Yikes! You can't please everyone I guess. A couple weeks ago readers were practically begging for full RSS with ads.

I think if we were to implement full RSS with ads, we'd have two feeds: one would be the standard partial feed and the other would be the full one with ads. Would that be okay?

Dave, you're concentrating on short-term revenue. It sounds as though you're having a similar mental block to the RIAA, seeing a primary revenue stream decline and freaking out when the numbers really indicate a strong demand for the product. Those RSS subscribers are subscribers, not just people picking up a copy on the newsstand as they wait for a plane. They'll be coming back to read you even if you don't have a link from somewhere else.


The point of the experiment was to see if there was a short term benefit to full RSS. There wasn't, at least not that I could identify. If there had been, it would have been a no-brainer to stay with full RSS feeds.

But that certainly doesn't preclude a long-term application of full RSS, and indeed we are considering doing just such a thing. Unfortunately more people than just Greta and me need to be convinced.

(reply to #8) :-) Yes, that'd be great!

By Jan-Maarten (not verified) on 21 May 2007 #permalink

Randy, I don't think so, it just looks that way because April 23-29 was an anomalously high week. The previous several weeks were similar to April 30-May 6. Here's a graph of our traffic going back to March 26.

"we'd have two feeds: one would be the standard partial feed and the other would be the full one with ads"

I rarely read partial RSS. You can monetize large RSS subscriptions via FeedBurner ads and others.

I like the compromise, two feeds approach.

I stopped reading your blog because of partial feeds and I don't read others with partial feeds. Today I saw a comment to this article on Scoble and thought I'd check out what was going on. A partial with an attention grabbing paragraph doesn't cut it. I want to read as much as I want right then and then stop, not as much as you want me to read. There isn't time to wait for another page load, that's why I use an aggregator. YMMV. Unfortunately, it seems your focus is on profits while I'm interested in the exchange of ideas.

Unfortunately, it seems your focus is on profits while I'm interested in the exchange of ideas.

Yes, I'd like to make money as a writer. Arguably others would like for me to make money too, since that increases my ability to keep "exchanging ideas." At some point, if I don't make enough, I'll probably need to do something else for a living.

What we're working on is a way to make money while not limiting the exchange.

I came to the site to view the research articles that I heard about on the CogDaily podcast. There have been no new podcasts since January - so there is nothing bringing me to the site. I love the articles and your brief synopsis on the research - it lets me know what is current in brain research without spending hours searching the internet or library. These allow me to go directly to the source, which leads me to other resources. When I am here to read the articles, I read other posts and some of the blogs and comments. Without the podcasts as a reminder of this great resource, I forget about the website.

Interesting to see this experiment and the resulting discussion. My first thought is that a week is not enough time to test for a real change in the page views or click-throughs. Even still, I'd be curious to check back in a month: what do the statistics above look like for a two month window. Is there an obvious dip during the experimental week? Or is it within the noise of everything else that's happening at your website?

On the full v partial debate, I lean towards full. And that means I lean towards not visiting a website, except to comment or (rarely) dig back into the archives. However, since I actually read full-text feeds, I am much more likely to reference articles from full text feeds (because I read them). Unless the partial text is _exactly_ something I need, I usually don't stay subscribed for very long. Which means that website never gets my attention unless I get drawn by someone else, which is what happened today.

I hate partial feeds. HATE them.

hey, we need to do some proper econometrics here!

a) what is the trend in the two numbers before the change?

b) what are the results after you introduced the change, adjusting for the trends?

Plus one note: RSS subscribers have a big advantage. If you stop blogging for a week your page subscribers will fall. Your RSS subscribers will not! A good RSS subscriber base is thus much more stable than a mass of users who are visiting now but who might focus somewhere else if you are not so active for a period...