Bloglines subscription stats? Just check your httpd access logs.
We were chatting about tracking more accurate feed subscription stats after work the other night, and I made a mental note to poke around my httpd server access logs (thank you Dreamhost for making that so easy). So I’m paging through yesterdays 20MB access log, when the user-agent string for the Bloglines bot catches my eye:
Bloglines/3.0-rho (http://www.bloglines.com; 7 subscribers)
Whut? That prized information, the number of people Bloglines is redistributing my feed to is right there! Ok so it’s not exactly an API. And when I started looking through the user-agent strings, I discovered that there was a wide variance in how these web based aggregators reported their subscribers. Here’s a sampling:
NewsGatorOnline/2.0 (http://www.newsgator.com; 5 subscribers) Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1; Rojo 1.0; http://www.rojo.com/corporate/help/agg/; Aggregating on behalf of 3 subscriber(s) online at http://www.rojo.com/?feed-id=2392387) Gecko/20021130 kinjabot (http://www.kinja.com; 1 Readers) YahooFeedSeeker/2.0 (compatible; Mozilla 4.0; MSIE 5.5; http://publisher.yahoo.com/rssguide; users 1; views 1) FeedLounge (http://dl0.feedlounge.com:9999/), 1 subscribers, next refresh in approx. 28800 seconds LiveJournal.com (webmaster@livejournal.com; for http://www.livejournal.com/users/jstnsmnia_blog/; 1 readers) AttensaOnline/1.0 (http://www.attensa.com; 1 subscribers)
Wouldn’t it be nice if there was a standard? Something like:
X subscribers;
The rest (as far as I could tell), were just regular individual requestors. Or web-based aggregators that hadn’t jumped onto the custom user-agent bandwagon yet, *cough* Google *cough*.
Feedfetcher-Google; (+http://www.google.com/feedfetcher.html)
With a little bit of moxie (grepped the access logs, regular expression formatting in EditPlus, analysis in Microsoft Access) I can say about 190 people subscribe to my feeds. Which is surprising to me. Here’s how it breaks down visually for my info-porn fans.
Apparently Google Desktop is a force to be reckoned with. I had more of these in my logs (coming from unique IP addresses) than any other aggregator:
Mozilla/4.0 (compatible; Google Desktop)
And I had no idea Google Desktop included a feed reader.
Google Desktop is not really a subscription and shouldn’t even be counted as anything. Google Desktop will hang onto rss feeds it comes across and keep them updated until they are updated (by the user going to more webpages) to some other feed. This happens even if the user doesn’t have that option displayed. It’s really only temporary and unintentional, and more than likely people with google desktop are not even reading what it’s ‘subscribed’ to automatically.
Bjorn, thanks for the heads up. I’ll have to give Google Desktop a test drive to confirm sometime.
how do u make a pie chart
Joe, in this case, I used Microsoft Excel to create the pie chart and then I took a screenshot (PrtSc) which I pasted into an image editing program to crop for my blog.