Realistic blog usability

July 16, 2010

Many moons ago, Jacob Nielsen wrote an entry about 'blog usability', which I reacted to at the time. But as I noted back then, Nielsen's advice seemed aimed mostly at what I called 'commercial' blogging, and ever since then I have had simmering in my mind the idea that it misses the true usability issues for typical blogs (then and now). Today I feel like letting things bubble over and finally writing down my thoughts.

To start, let's ask two important questions: how do visitors come to your blog, and what do you want them to do once they've arrived?

Until you get famous, I think that most of your visitors will come to you either from search engines or (these days) from links shared through the 'social web' (Twitter, Facebook, Reddit, etc). In both cases they're mostly going to go to specific entries, not any of your index pages (front page, category pages, etc). My answer for the second question is that once people have arrived at an entry, you want them to stick around and browse further in your blog; ideally you'd like them to become regular readers.

(Back when I first had this thought, search engines were by far the dominant traffic source you could expect. The growth of the social web may have changed that, especially if you're writing the sort of blog entries that attract links. WanderingThoughts still seems to get most of this sort of traffic from search engines, but that's a function of the sort of entries I write.)

These answers have some consequences. First, you really want people to get what they came for, especially if they came from a search engine. At one level this is out of your control, but at another level you want to make the search engines prefer only your most stable and most specific pages, ie the entries themselves. A visitor who is directed to an ever-changing index page is unlikely to find what they're looking for and if that happens, they're unlikely to look any further at your blog; they'll just hit Back (or close the tab) and go somewhere else.

Once a visitor has found what they're looking for on that initial page, what you need next is things that encourage them to explore outwards into the rest of your blog. There are a number of potential hooks here; I think that the most powerful one is a 'related posts' feature, so that your visitor can continue on the topic of the entry that they just finished reading and are (presumably) already interested in.

(A 'greatest hits' feature is attractive, but it sort of requires the visitor to be interested in your writing in general instead of just the specific topic that the entry they landed on is about.)

One way to encourage exploration is to make it less uncertain and risky, specifically by giving people more of an idea what they'll get when they click on a link. Thus, I feel that it's important to have good entry titles and to put them in links on other pages; for example, links to the next and previous entries are (much) more attractive if they show the entry titles than if they're just labeled 'next' and 'previous'.

(As a corollary, category and tag links are less attractive than a 'related entries' set of links that actually shows the titles of those related entries. Even if the actual category page shows more information, people have to do that initial click based on very little but trust.)

This does make me think that the 'many entries on one archive page' model of blog archives may be less insane than it looks, since it lets visitors browse further with essentially zero effort; this is probably especially attractive if you archive things by topic instead of by date. However, to make this really work I think you'd want some JavaScript magic to automatically position the page at the entry that best matches the search terms that search engine visitors used. Otherwise you risk presenting searchers with a large mass of text that isn't really what they were looking for, causing them to bail out on you.

Astute observers will notice that WanderingThoughts falls down on many of these measures. One of the reasons for that is that I had all of these thoughts only after I wrote DWiki, the engine behind WanderingThoughts (and I wrote it for a different purpose, to boot). If I was building a file-based blog engine today, it would look quite different.

Written on 16 July 2010.
« Making the Linux kernel shut up about segfaulting user programs
More building blocks of my environment: tkrxterm, tkssh, and pyhosts »

Page tools: View Source, Add Comment.
Login: Password:
Atom Syndication: Recent Comments.

Last modified: Fri Jul 16 00:04:30 2010
This dinky wiki is brought to you by the Insane Hackers Guild, Python sub-branch.