Tuesday Sep 11, 2007

Jakob Nielsen talks about blogging, search, and documentation

Jen McGinn is an interaction designer in xDesign who is working to improve the user experience with the Java Enterprise System installers. She has an MS is Human Factors in Information Design and works out of Sun's campus in Massachusetts.

Jakob Nielsen, formerly a Distinguished Engineer here at Sun, visited with some of the folks on our Menlo Park campus today, to give a brown bag talk on his July 9th AlertBox article "Write Articles, Not Blog Postings".

He started off the talk by saying that his analysis was based on Peter Pirolli's 1999 work with Stuart Card, Information Foraging.

He talks in what I know as terms from economics — the opportunity cost of pursuing a line of action; the diminishing returns (utility function) on each unit of information gained; and the cost of obtaining that information (cost function). He goes on to talk about an example of the two Australian scientists who discovered that ulcers were actually caused by a virus. When they won the Nobel prize for their work, it was reported in every newspaper around the world. After reading about the finding in a newspaper, a physician might want to read up on the authoritative source. What kind of information is your reader looking for? Authoritative sources or quick information?

He suggests that blogs are "stream of consciousness" where their authors are just talking about thoughts off the tops of their heads. He sees the value of blogs is to "pump out" new content. An audience member (Richard?) compares his own experiences with blogs and bloggers as being more similar to newspaper columns and columnists — not that they are necessarily providing authoritative info, but that bloggers can connect the reader to some content that the reader is interested in, providing both entertainment and the ability to navigate the web. The audience member also notes the collaborative nature of blogs — that people can easily communicate and participate in new and developing works — and that they have the timeliness that traditional authoritative documents can't have. Nielsen agrees, but says that the linear structure of blogs does not lend itself to good usability. Another audience member suggests using a table of contents, and another suggests tag clouds. Nielsen suggests that those are not conducive to problem solving, because they do not highlight what is most important or salient.

Nielsen asserts that search engines are trying to answer 2 questions: what is this about, and how important is it (page or relevance ranking). But what we really are looking for is a measure of usefulness — how will this search hit help me solve my problem.

He goes on to say that blogs lack editorial style and polish, but that there is a cost to that effort (back to utility and cost functions), and that extra cost and effort would not allow for blog content to be produced and released quickly. Compiling, summarizing, and concluding are three things that are not typically included in blog posts. But later, he says that blogs written by multiple people can be more valuable, because it's hard to create valuable content every day, but easier with multiple contributors.

Another listener suggests that different audiences and tasks are better suited to be answered by informal and formal communications. For example, when he wears his developer hat, he is interested in knowing what other people are doing, but when he's a system admin, he wants official documentation. Paul laughingly suggests that that's a tech writing 101 concept, to perform an audience analysis. Janice Gelb suggests that having both product docs and blogs means that we have two separate places for people to look for information. Nielsen suggests that the two types of content ought to be better integrated by referring to one another. This information sharing between silos also helps to educate the search engines.

Role-based personalization can be more helpful than customization, because the information is tailored for a person based on known categories of attributes, versus the user having to do something to customize their system. If only we could guess what people need and give them the three things that are the most useful, they would be ecstatic. His counter-example is an online book-seller, which makes lots of personal recommendations, but almost never recommends anything that anyone really wants. What does work well, though, is its cross-selling mechanism, letting you know about the books bought by other people who bought that book, because it's built on lots and lots of people's past behavior.

Hey, past behavior is a good indicator of future behavior :) That's a tenet of user research, which is why we are creating personas based on customers' behavior. But that's a topic for another day ...


xDesign is a software user experience design group at Sun.
Follow us on Twitter : Flickr : Blog (see feeds below)


« July 2016