What is the most popular social bookmarking site? There are many metrics to measure this, and one of them is submissions of links. This website is the home of the Socializer, a multi-social bookmarking service. Using data collected from visitor clicks using the Socializer, I ranked the popularity of the services it lists.

Methodology

Hundreds of sites use the Socializer for their social bookmarking needs. For each page, they create a submission link to the Socializer. When visitors arrive on the Socializer submission page, they are presented with 47 links, each being a specially formatted URL for submitting a link to one social bookmarking service. The 47 links include the most popular services like digg, del.icio.us, and reddit, and many niche and less popular services. The services are listed alphabetically in a rectangular grid, starting with Backflip on the top left corner. Here is an example.

When a visitor clicks on a submission link, the click is logged using the AJAX Link Tracker with the help of a server-side script. Because it is using AJAX, the visitor's submission is not interrupted; i.e., the data collection does not interfere with the actual bookmarking activity. Since the Socializer page contains other links than just the submission links, clicks on those are also recorded. Example of non-submission links are navigation and links to external sites that host plugins for getting the Socializer to work with Wordpress, MovableType, and others.

A few weeks ago, I reset the click logs and started counting from scratch. As of 11PM last night, I had collected 21835 clicks. Overall, these clicks represented 70% of all clicks on the Socializer's page.

Results

Digg was the service most submitted to. Digg collected 1220 clicks (5.6%) during this time. This represents a very wide lead over Netscape, the second most popular service, which collected 1009 clicks. In third and fourth positions are Yahoo!'s two properties, MyWeb 2.0 with 981 clicks and del.icio.us with 942 clicks.

The rest of the data is presented in the table below and graphically.

(Click for full size version.)

Analysis

There are many points to make here:

• Digg is very popular indeed. It has a full 21% gap over the second place service, Netscape
• Speaking of Netscape, everyone yawned when they moved to their new Digg-like service. They seem to be doing quite well to me!
• The data agree with a recent post by Rand Fishkin. He talked about links gained from social sites, and Netscape is up there with Digg, del.icio.us, and reddit. That's another metric of popularity of social sites, and it's good to see some agreement.
• There is a very tight pack trailing Digg composed of Netscape and the two Yahoo! sites del.icio.us and MyWeb 2.0. The difference is minimal, but real as far as I can tell: For the past five days, I've been checking the data, and the rankings didn't change, but they all gained clicks (as expected). It will be interesting to watch the top four in the future.
• But, but, but: although del.icio.us and MyWeb 2.0 are distinct services, together they tell of a very powerful Yahoo! in the social bookmarking market: Combined, they garnered 1923 clicks, which is 8.8% of the collected clicks. Yahoo! is the true leader here.
• Notice the green trend line I've plotted on the graph. It's a best fit logarithmic trend, and the fit is really good. This harkens to the idea of a 'long tail' where we observe very few categories (in this case social bookmarking services) being the most popular followed by a long tail of categories of ever decreasing popularity. This is interesting and warrants further study.
• We have to stop and think about toolbars. The key service here is StumbleUpon which shows a very paltry performance. However, most SU submissions are probably done using their toolbar that users install, and so we expect to see very few people using their submission URLs.
• One service I'm surprised to see perform so well is PlugIM. I've mentioned it several times in previous blog posts as it's relevant to internet marketing. I asked the creator of PlugIM to give me exact details, and he said that PlugIM officially launched September 1st of 2006, a mere five months ago. In this light, they're doing really well!
• The performance of Backflip amazes me again. Last time I did this league table, Backflip also performed very well, but frankly, I don't hear much about it on the web! One suggestion is that since it is the top left link, it is the first link visitors see and click it anyway. This is a bit like putting ads in the top left triangle area of a web page. I accept that this is a real possibility, but I'll study it more before making judgement.
• We have to note again how the data was collected, namely, using Javascript. Although not all people have Javascript enabled, there is no reason to believe that those that do and those that don't behave differently from each other. I'm very happy to accept that users browsing with Javascript enabled are a representative sample of the Socializer's users in general.
• Finally, we have to look at the demographics of the Socializer's users. The sites that use the Socializer are diverse and target very different markets. The markets include technical sites, adult sites, commercial sites, and charities. I have no reason to believe that the sites using the Socializer are not representative of websites in general. The largest category of sites is blogs, though, but I don't know the exact percentage.

Future Work

Well, this is the second time I do this ranking, and the results are very interesting. In a few months time, I'll quietly reset the counters again and do the experiment again.

Also, there is a new version of the Socializer in development. It will have many new features, some aimed at gaining better understanding of the how users use social bookmarking services. The data will be published like this routinely.

In the meantime, I've partnered with SEO Refugee Forums using a customized Socializer, and I'm looking for other partners too. No data has been collected, but it will be interesting to see if the global usage pattern holds for individual sites or not. Should be fun

Any questions or comments, please comment below or email me.