Pure Danger Tech


Blog Tornado

21 Mar 2007

As I’ve started posting more frequently in the last few months, I’ve seen my blog traffic increase substantially. In particular, I’ve watched with fascination how a post will get picked up and if it’s interesting will get sucked into a “blog tornado” that spirals up through a series of blog and news aggregator sites. I can clearly watch it happening if I go back through referrer logs on my site. They vary as to how much they are automatic vs human-submitted, filtered vs unfiltered, time lag from initial post, and audience size. My blog is mostly about Java so my examples below are all in terms of the Java blog community. I’m sure similar examples exist for other topics as well.

Seed aggregators

Seed aggregators are automatically aggregated (through registered RSS feeds), totally unfiltered and high traffic. They are the firehose of aggregators and typically have a lot of noise. Your entry will be posted to a “new” page and due to the traffic, your entry will quickly (hours, minutes, or days depending on the site) get scrolled off the “new” page. However, most aggregator sites track articles by clicks and/or views and automatically collect popular entries on some kind of “hot” or “rising” page.

For Java-related blogs, the first and most direct initial dispersal for me occurs through java.blogs where my blog’s feed is registered. As far as I can tell, this is the most open Java blog aggregator site on the net as anyone with a Java-related feed can register here. java.blogs has a front page with incoming posts and automatically tracks clicks and views to generate a hot entries page. java.blogs alone can account for anywhere from 10s to 100s of hits on a new entry. The quality of your title and first few sentences (and ultimately the entire post) has a big impact on the traffic. The time and day when you post also can have a surprisingly large impact on how many people see it, as the volume affects how long you are on the “new” page. You could probably determine an optimal time to post, but I’ve never tried.

Most of the other good seed aggregators out there are not open – they typically require you to use their service or have some special affiliation. Some of the best are JRoller at JavaLobby (requires you to have your blog hosted there), java.net (for blogs hosted there), and Planet JDK (for those working on the JDK).

To some degree, you could see Technorati as the widest reach as a seed aggregator, although it is so broad that I very rarely see any traffic from it or the other blog indexes. Really, at that scale it bleeds into search (mentioned below).

Filtered Aggregators

After the seed aggregator creates some dispersal, a bunch of people will see the entry. If it sucks, then that will pretty much be the end of it. But if not, then some people that see your entry will like it. And the first thing they are likely to do is post it to a filtered aggregator. Filtered aggregators have links submitted by people, so they typically have lower volume, less noise, and higher quality links. They also often have feedback mechanisms to track hot links and elevate them to prominence.

My favorite filtered aggregator to use (and one that drives a fair amount of traffic) is dzone. I really love their presentation, filtering, etc. I used to see my entries get submitted by others on dzone but now for entries I write that I think might have some broader interest, I usually just submit them myself.

In the Java world, you also can’t escape from the influence of Erik’s Linkblog, one of the oldest filtered link collections I’m aware of. Usually, if I get picked up on the linkblog, it comes a day or two after the entry gets popular on dzone or some other site. The initial wave of attention is just dying down when the linkblog is published and a secondary cycle of interest starts. The linkblog also gets republished on some other web sites like planetjava.com and goes out through email, further increasing the reach.

I should also mention the mega-aggregators like digg and reddit. These are also filtered aggregators but they tend to reach a very wide audience and due to the voting systems, only items with very broad applicability survive the initial posting window. I’ve had a couple things picked up on the programming subreddit at reddit but I generally don’t write anything with broad enough interest that those posts would last very long. I’ve also had some traffic by services like StumbleUpon, although that seems like a little more randomized than something like reddit or digg.

A special kind of filtered aggregator is a site devoted to a particular project, language, etc, often filtered by just one or a few people. One of my articles about ANTLR was listed on the ANTLR article page (and home page) and that alone has driven a steady stream of traffic for months.

Bookmarking services are similar to filtered aggregators (in that they are human-selected) but differ really in their intention. With filtered aggregators, your intention is primarily to share or promote a link with others. With bookmarking services, your intention is usually first to remember the link for yourself and only secondarily to recommend it to others. Or at least that’s how I use it. So, if I write entries that are very focused and descriptive, then I will sometimes link it up on deli.cio.us. I do see a trickle of constant traffic from deli.cio.us tags, although mostly through other people’s tags, not mine.

If your entry has been well-linked by others, it will start to get picked up by search engines like Google or Yahoo. There’s plenty of sites out there on search-engine optimization, so I won’t touch that. Can’t say I’ve really done anything with it other than just pinging the right blog RPC sites at Yahoo, Google, and Technorati.

These happen at a fair lag, usually days or weeks behind the aggregators. But they are important as they drive long-term traffic.

Other Bloggers

Ultimately, the highest-quality and most interesting backlinks come from other bloggers referring to what you’ve written. Sometimes these occur immediately as people either reading your feed directly or seeing it via a filtered aggregator react to what you’ve written. Sometimes they happen at great remove as someone searches or finds your entry through some other means. In any case, backlinks cause search engines to rate your site higher, making it more likely that they get picked up on search queries.

If you want to increase your interaction with other bloggers, you can comment on their posts (leaving a link back to your site), write blogs referring to them and trackback to your site, or add their blog to your blogroll. If someone links to you and has an interesting post, you can push them into the blog tornado by posting their entry on a filtered aggregator. That will often give you a secondary effect as people read the new article and follow the link to your site.


I find the way all this stuff works kind of fascinating. I like to promote my site because it’s more fun to write when people respond to it. I also find it kind of fun to see if you can hack the system itself (go figure – I’m a programmer). I think the feedback systems that are emerging in the filtering aggregators are particularly interesting and will continue to evolve into (hopefully) ways to see better links with less noise over time. It will also be interesting to see how blogs themselves evolve in tandem with these sorts of tools.