Hacker News Reader

Comment by susam | Hacker News Reader

I provide an RSS feed for my blog. From my access.log, I see that there is a decent number of subscribers to my feed. Here is the data I could pull from it:

  Feedly           91 subscribers
  Inoreader        16 subscribers
  Feedbin           6 subscribers
  NewsBlur          4 subscribers
  The Old Reader    3 subscribers
  BazQux            2 subscribers
  WordPress         2 subscribers
  -------------------------------
  Total           124 subscribers

I get about 4000-6000 hits to my website on a normal day. It increases by 5 to 10 times if a post hits the HN front page. When I publish new post, I see from the HTTP referer information in the logs that about 30 or so hits come from users who find a link to my post in their feed reader. These are very small numbers but they are good enough to keep the feed alive.

View on HN

I came here to say exactly this: my numbers are similar to yours (and from headers only, like you[1]), and a decent proportion of my views come from feed clients. My blog generator takes care of the feed for me; I haven't thought about managing it in years.

[1]: https://www.yossarian.net/snippets#vbnla

I have about the same distribution of RSS clients on my blog as well. Feedly seems to be the largest, with about 10x reported subscribers than their competitors. Newsblur, The Old Reader and Inoreader are all about the same, but far behind. Then there's a sea of various self-hosted readers like the TinyTinyRSS and desktop programs.

Note that these are all self-reported numbers from the user agent string - as far as I can tell, Feedly might be inflating their numbers to give an impression of a large user base and to appear on top of lists like this.

How do you identify Feedly subscribers? Isn't there a single request from Feedly that it uses for all of its users?

Here is what I see in my access logs:

  [30/Jan/2021:12:58:03 +0000] "GET /blog/rss.xml HTTP/1.1" 200 22757 "-" "Feedly/1.0 (+http://www.feedly.com/fetcher.html; 91 subscribers; like FeedFetcher-Google)"
  [31/Jan/2021:12:58:04 +0000] "GET /blog/rss.xml HTTP/1.1" 200 22757 "-" "Feedly/1.0 (+http://www.feedly.com/fetcher.html; 91 subscribers; like FeedFetcher-Google)"
  [01/Feb/2021:12:58:13 +0000] "GET /blog/rss.xml HTTP/1.1" 200 22757 "-" "Feedly/1.0 (+http://www.feedly.com/fetcher.html; 91 subscribers; like FeedFetcher-Google)"
  [02/Feb/2021:12:58:19 +0000] "GET /blog/rss.xml HTTP/1.1" 200 22757 "-" "Feedly/1.0 (+http://www.feedly.com/fetcher.html; 91 subscribers; like FeedFetcher-Google)"
  [03/Feb/2021:12:58:30 +0000] "GET /blog/rss.xml HTTP/1.1" 200 22757 "-" "Feedly/1.0 (+http://www.feedly.com/fetcher.html; 91 subscribers; like FeedFetcher-Google)"

So feedly tells you how many subscribers they have when they hit your feed. That's neat of them.

Not only Feedly does this. Most of those feed readers do it.

In fact, Feedly was the only large centralized feed reader that didn't do subscriber count reporting via UA string for the longest time. They made you do a GET request on some API endpoint if you wanted to get your subscriber count.

They implemented it some time back, but it's still kind of buggy - at least for my blog. Two or three different IPs periodically do a GET on my feed URL with the Feedly UA and they all report different subscriber counts.

It might not be Feedly?

Seems unlikely, since they are all from the same subnet: x.x.x.25, x.x.x.26, x.x.x.27

It is unfortunate that it's so tough to track RSS viewers. I'm always tempted to convert the feed to cut-back version just so I can see what posts are actually popular.

An idea I had is to just embed a pixel into the item body. You can self host it too: just analyze the apache logs (or whatever you have) for requests to that pixel.

Feedburner was always the way. Then Google bought it and I’m not sure what happened next

Almost literally nothing other than the domain changing. Don't remind them they have it, they might shut it down :P

Something I'd be curious about: how many of the hits on a new post are by RSS readers before the post hits HN vs. after? I know RSS is the most reliable way for me to notice a new post on blogs that update infrequently, and I wonder how many HN submissions come from people seeing a new post in their feed readers.

What is the topic of your blog?

I don't write very often but when I do it is usually tech or math. I didn't link to it here because it might look like self-promotion. In case you are interested to take a look, please see my profile for the link.

Hey! I read some of your articles a couple of days ago :) I liked your blog.

Thank you! I am glad you liked my blog. :)

Reply on news.ycombinator.com