Why Digg Falsely Increases your RSS Feed Subscriber Count
Posted by Jon Lee in Site stuff, Web Development, tags: caching, digg, Feedburner, firefox, google, Google-Web-Accelerator, prefetch, RSSLast week, my post with the latest Starcraft 2 gameplay videos made it to the front page of the digg video section and received over 450 diggs. As a result, over 10,000 visitors poured in and the next day, my Feedburner count was at 266 (more than double the previous day’s count of 130). A couple days later, the count dropped back to a normal level of ~140.
Here are several theories to explain why this happened:
Theory 1: Real subscribers
People clicked the digg article, liked what they read and subscribed to my feed. The following days, I made a post about ReviewMe and the launch of my ShowoffRankings site and these people that had newly subscribed found out my blog isn’t really about Starcraft and unsubscribed as a result.
I’m sure a few people may have done this but it doesn’t explain the temporary spike. Proof? 3 months ago, I had a post that made it to the first position of the digg top 10 with almost 4,500 diggs and received close to 100,000 hits. The next day, my Feedburner count went from a measly 6 subscribers to an all-star level of 3826 subscribers. Over the next few days it quickly dropped down to a humble 30 subscribers. I refuse to believe that almost 4000 people subscribed to my blog only to unsubscribe a few days later because they didn’t like my content!
Theory 2: Firefox Prefetching
One theory I have heard is that Firefox prefetches links to improve browsing experience and by prefetching your RSS feed, you get a temporary “subscriber” without having them actually click on your feed. If you take a look at the subscriber information before and after the digg, you can easily see that the increase of “Firefox Live Bookmarks” subscribers almost accounts for the entire subscriber spike.
So the evidence seems to support this argument, but wait – almost half the visitors to this site use Firefox, so wouldn’t that mean with 10,000 hits I should have 5,000 extra subscribers and not just an extra 150? Well by default, in Firefox configuration, the “network.prefetch-next” value is set to true. This however, does not mean that Firefox is prefetching links for you — it means that Firefox will prefetch links that the website tells it to prefetch using an appropriate HTML declaration.
Theory #3: Firefox Prefetching (with add-on)
The previous theory seemed to explain it pretty well except for the fact that not all users are triggering the prefetch. The answer is in the individual Firefox installation – namely, whether or not a user is using the Fasterfox extension or has Google’s Web Accelerator installed. Both of these applications will cause prefetching to occur without having the website to explicitly tell it to do so. Of the 10,000 visitors, if about 1% of them have either of these installed, it would completely account for the increase for the spike.
Conclusion
Now none of these theories are proven but I think the third one sounds pretty reasonable. I’m doing a little experiment right now to test out the theory so if you want to help me out, you can digg yesterday’s article about creative outdoor ads in order to push it to the front page and bring in a load of visitors for me to test with 
I’d also be interested if anyone else has any theories or knows for sure why this phenomenon happens. If you do, leave a comment!
Popularity: 3% [?]
Entries (RSS)
ahhhhhh really interesting … so thats why my feeds keep going up and down on subscribers!!!!! im def helping u!!!!!
Interesting. Personally, I haven’t experienced any increase in subscriber base myself. I wrote about it here:
http://www.terminally-incoherent.com/blog/2007/05/28/digg-effect-aftermath/
My subscriber graph was amazingly stable during the week when I got dugg – no spikes, no nothing. You’d think that I would get some of that Live bookmark and prefetch juice. But no.
Not sure what to make of this. I guess the main difference here was the demographic of digg users. You got dugg mostly by gamers who by definition are mostly windows users. I got hit by linux users who – especially those who like to live on the command line.
If anything I’d think that Linux users would be more likely to increase RSS feeds. One thing I noted about your blog is that your single post pages don’t contain any links to your feedburner RSS feed.. so there’s nothing to prefetch!
There is an “entries RSS” at the bottom of your page but that doesn’t link to your feedburner RSS feed!
Ah, good point. I haven’t thought about that.
Thanks!
[...] this week I wrote a post discussing why a huge spike in traffic will cause an abnormal spike in RSS readership. All the theories I suggested are just that, theories. So I set out to do a little experiment to [...]
I’ve never had an RSS increase due to being Dugg either. Weird.
I look forward to the findings of your research/experiments.
Findings are at: http://www.jonlee.ca/an-evil-way-to-increase-rss-stats/ but I see you’ve found it already
[...] Why Digg Falsely Increases your RSS Feed Subscriber Count | jon lee dot … A fun web development blog documenting my life and experiences … Why Digg Falsely Increases your RSS Feed Subscriber Count. Clever and Creative Outdoor Ads … [...]
o… dear, never explore about this “digg” subject that deep…. well I just use digg as it was.. to “spam”… LOL.. just kidding thou, but it is really useful to bring out our longtail keyword spread to the world… but I would sure you, don’t depend on the quality of the traffic, because it is bad… not good for adsense conversion rate…. so mind the risk since the BIG G is become more maniac than ever… you don’t want to loose you adsense account do you?
i dont really know much. but maybe they get all excited about the dig, subscribe to your feed then yes, they find out at your site that its not the same thing. so they unsubscribe.
I know this is an oldish article, but you could test out theory #2 by running Wireshark. I’ll give it a go sometime…