Jakob Nielsen Comes to the Sun Microsystems Editorial Forum

If a week is a long time in politics, a fortnight must be an eternity in the blogosphere. Consequently, it seems a little odd to be writing a blog post about an event that occurred two weeks ago, especially as, once again, I find myself undone by my technological conservatism: While I was furiously scribbling with pen on paper, Jennifer McGinn was already typing up her blog and linking to relevant content.

As I have mentioned before, one of my extracurricular activities is to chair the Sun Microsystems Editorial Forum. A couple of weeks ago, Jakob Nielsen came to speak to a combined session of the Editorial Forum and Sun's Software Information Products Group (IPG). Jakob is a former Distinguished Engineer at Sun and one of the leading experts on web usability issues, including documentation types and their usage on the web.

Jakob's talk was about recent research he has done on the relative value of writing articles compared with writing blog postings. Jakob's talk was also a continuation of the conversation with the Sun technical publications community about how to move Sun's technical documentation into the new millennium.

People hunt for information like wild animals hunt for food

Jakob began his talk by introducing the concept of information foraging, which was developed at the Palo Alto Research Center (previously Xerox PARC) by Stuart Card, Peter Pirolli, and colleagues. Information foraging is based on the use of mathematical models to describe the behavior of people collecting information on the web. These models suggest that people hunt for information on the web like wild animals hunt for food: In both situations, the hunters are looking for the maximum return on their efforts.

  • For wolves hunting in the forest, the maximum return on their efforts comes from hunting small rabbits, which are easier to catch than large rabbits.
  • For people hunting for information on the web, the maximum return on their efforts comes from reading short articles, which are quicker to read than long articles.

So, it would appear that short is generally better, and that perhaps, short blog postings are better than in-depth articles.

One exception to the rule that short is better

When people care enough about a topic to require in-depth information about the topic, the value of long articles increases. For example, the value of long articles increases for people such as:

  • Doctors who are looking for the complete list of foods that might kill a patient with a particular condition
  • System administrators who are trying to find out how to get 400 impatient users back online

Time that doctors or system administrators spend reading long articles from start to finish is repaid with interest because they don't need to spend time searching and retrieving a large number of short articles to get all the information that they need.

Structure aids search

Jakob suggested that search engines are reasonable at, but not brilliant at, conveying the following information about a web page:

  • What the web page is about
  • How important the web page might be to a reader

However, search engines are less good at conveying the usefulness of a web page, as users who were searching for information about the need to speed up page downloads but found only information about the Need for Speed video game will testify.

Good information architecture, coupled with writing that is optimized for the web, helps search engines convey the usefulness of a web page. An inverted pyramid structure that puts the most important information at the top of the information hierarchy and at the top of each page helps users quickly determine whether a page that is returned by a web search is actually useful to them.

A blog is a stream of consciousness

Jakob described blogs as a stream of consciousness: What was posted yesterday is forgotten today and by next week will have been pushed down by newer postings into the archive, where it will be lost forever.

Richard Friedman said that blogs are most valuable as reportage and as a source of referrals to more in-depth information elsewhere.

Geertjan Wielenga suggested that the problem of retrieving old information from a blog archive could be addressed through providing a table of contents for the topics in the blog. The blog as book, if you will.

John Domenichini (I think) suggested that a tag cloud could also address the problem of retrieving old information from a blog archive.

Another person wondered whether reliance by a corporation on a strong presence by an individual in the blogosphere could put the corporation's brand at risk. If the person and the brand become strongly associated and the person then leaves the corporation, might not the brand be diminished as a result?

Blog postings are most useful as newsletters

In the context of technical documentation, blog postings are most useful as newsletters. To remain effective as newsletters, content gardening of blog postings is required. However, the need for currency in blog postings leads to a shortage of links in blog postings, which militates against content gardening.

Good product support helps sell products

Jakob observed that good product support helps sell products by building long-term relationships between suppliers and customers. A critical component of that support is product documentation that customers can rely on.

In product documentation, task orientation gives better search results because users are already familiar with the words that describe genuine user tasks.

Wikipedia is not as good as it is made out to be

Jakob stated his opinion that Wikipedia is not as good as it is made out to be. He acknowledged that Wikipedia does have the open-source benefit of easy correction of factual mistakes by a large community of contributors.

However, according to Jakob, Wikipedia does suffer from the following shortcomings:

  • The combination of poor writing skills and a lack of editorial control often leads to rambling articles.
  • Contributions to Wikipedia reflect the knowledge and interests of the contributor.

As Jakob made his observations, I was reminded of similar reservations that Nicholas G. Carr expressed in his article The Ignorance of Crowds:

Wikipedia's problems seem to stem from the fact that the encyclopedia lacks the kind of strong central authority that exerts quality control over the work of the Linux crowd. The contributions of Wikipedia's volunteers go directly into the product without passing through any editorial filter. The process is more democratic, but the quality of the product suffers.

I also saw an analogy with contributions from users to documentation for software products: Users write about what they know. So, some common use cases might not be covered in a documentation set that is comprised solely of contributions from users.

Business to business suppliers should avoid the latest fads

Jakob cautioned business to business (B2B) suppliers such as Sun against the latest fads. The B2B sales process is usually much longer and more complicated than the business to consumer (B2C) process, and purchasing decisions are not normally taken by one person alone.

Streaming media have their place, but they shouldn't be ubiquitous

Jakob acknowledged that streaming media can be useful in some situations, for example:

  • For overviews
  • Where control over the linear flow of information is needed
  • Where sound and movement convey the information much more effectively than words, for example, in hardware maintenance procedures

However, imitations of streaming media include:

  • Poor search. Only the index tags that are assigned to an item streaming media can be searched, not the actual content.
  • Linear information flow. In text, users jump around (so to speak) because they don't read: they scan. Streaming media are unsuitable for learning where the learner is self-paced.

How users consume information will change, but not as much a we might think

David Lindt asked whether youths who are growing up with facebook, myspace.com, and other social networking sites have different expectations for online information from older generations. Might they not expect personality and right to reply in their online information? Jakob replied that when these youths have matured and are doing the types of jobs (say, system administration) that adults are doing now, they are likely to act in much the same way as adults do now.

Jakob did acknowledge that the 40–year olds of the future will probably consume information differently from 40–year olds now, but not as differently as we might think. He also cautioned that we cannot predict how people will behave in the future. We can only observe how they behave now.

Personality might belong in documentation sometimes

After the meeting broke up, I managed to speak to Jakob individually and to thank him for talking to the Sun writing community.

I also managed to ask the question I had not been able to ask during his talk: Does personality belong in documentation and other instructional information?

I cited the Virgin Atlantic Safety Video as a rare example of where the injection of personality actually increases the effectiveness of the information. This video uses cartoon animation, comic-book style characters, and humour to convey serious and important information. This combination certainly worked for me. The first time I saw this video, I was already an experienced flier with other airlines. However, the style and tone encouraged me to pay sufficient attention that I actually learned something I didn't already know.

Jakob and I concluded that the injection of personality was effective in this instance for the following reasons:

  • The information is being delivered to a captive audience with time on their hands.
  • As most people in the audience believe they are familiar with the information already, what might otherwise be a distraction actually causes people to pay attention.
  • The information is not intended to be acted upon when it is delivered, but remembered and (in the worst case) recalled if necessary.

Of course, the personality in this video is not the personality of the individuals who created it. Rather, this video reflects the corporate personality of Virgin Atlantic, which commissioned the video.


Thanks for this extensive review.

\*\*\* Geertjan Wielenga suggested that the problem of retrieving old information from a blog archive could be addressed through providing a table of contents for the topics in the blog. The blog as book, if you will. \*\*\*\*

That solution only covers the topic titles, not the content. If a blogger uses titles like "Cool New Feature!!!" or "Useful Script" then a ToC is useless. Plus, if a blogger is making daily (or even more frequent) entries, then the ToC would soon be unwieldy. Grouping listings by month or date would mean that readers would have to remember not only the topic title but approximately when it was posted. Not sure this works as an answer to enabling readers to find information. I could easily envision a scenario where a user remembers seeing an entry at some point with useful information but didn't take careful note of it because it wasn't relevant at the time they read it.

I was able to ask Jakob my question in the meeting, regarding whether the use of both blog postings and more persistent user information to provide documentation about a product meant that we were forcing users to look in two places and what should be done about integrating the information. His response was that the blog entry could provide a pointer to the persistent documentation.

Although I didn't pursue a follow-up in the meeting, I would have liked to have pointed out that this solution works in the direction blog->user doc but not in the opposite and more problematic direction, user doc->blog when blog entries document something that the persistent documentation doesn't cover. If blog entries are ephemeral and difficult to search but are covering aspects of a product that would be useful to users who don't necessarily access blogs, then this is a problem that needs to be solved.

Posted by JaniceG on September 25, 2007 at 03:16 PM PDT #

Thanks for this extensive recap, Paul. I wasn't able to attend Jakob's presentation, so this is great for me.

To deal with the issues of aging information and unauthenticated information (especially for B2B), we are going to have to establish a hierarchy or flow to help readers get from blogs to relevant docs, wikis, articles, etc. We've done that with release note, What's New, and docs. People understand the flow through those documents, and we have ingrain a similar understanding with these new sources. For awhile anyway, it is going to have to be explicit. Perhaps some of this can be handled with simple boilerplate references to key info sources and indications of the editorial control exercised over each of those sources.

To meet the need for brief, focused information, we need to tag all our information finely enough to be found. Blogs aren't the only things that are brief. Sun documentation has an abundance of highly focused information in brief chunks, but it often isn't possible to pull them out of the books they are contained in.

Posted by Jim Siwila on September 25, 2007 at 11:51 PM PDT #

You've done a superb job capturing and discussing the issues raised by Jakob.

I have a few points to add. After his talk, Jakob said to me that he couldn't urge us strongly enough to do usability testing of our new approaches. As a department, we haven't done many usability studies since Marney Beard and her usability testing group moved out of mpk17 years ago. Whether or not we have the time and resources to conduct tests, his point was a good one. We are moving into new forms of communication with our users and collaborators based on our sense of trends and the preferences of key internal stakeholders. Both legitimate motives for doing this, but neither gives us much to evaluate.

It would be useful and interesting, I believe, to see, for instance, how successfully a user finds technical information on a blog...finding user information in a current blog posting and then finding information in a posting dated months ago.

Or setting up a test to determine the effectiveness of putting out information on facebook. Dave Lindt brought up good points during the talk about whether teenage users of those sites might expect a different kind of writing style. Do we know what that style is? Maybe if we hire teenage interns, we will.

Will we even know and be able to respond quickly enough to changes in the demographics of sites like facebook?

On the matter of publishing information or advertising Netbeans Beer Bashes on facebook, there is a very interesting study of facebook by the Fabernovel Consulting firm at http://www.slideshare.net/popular

There are a few points there that should be taken as cautionary. One, the demographics of these sites change rapidly; I would say probably more rapidly than we are likely to keep up with unless we have resources dedicated to researching trends, countertrends, and disappearing trends.

This study shows that the demographics of facebook just in the last year have shifted from over 50% of its membership being in the 12 to 24 age range to over 50% in the 25 and older range, with the over 35 set being now the largest demographic segment. If we apply retro-causality, we might catch our intended youth movement developers of the future. If we don't, we'll be talking to ourselves by the time we get info out to facebook.

Two, another important point that they make is that "designing and spreading an application on Facebook requires a particular expertise that media companies seem to lack" and "Their applications fail to draw massive amount [sic] of daily active users." But before you object and claim "but IPG isn't stuffy like CBS News or the New York Times," their next point is "Even Yahoo!, a company that knows how to distribute media on the Web, had to turn to application giant RockYou to redesign its application and attract users."

There is more there on slide 22. All worth emblazoning on our collective strategic brains before we make the same mistakes we used to make when word processing became a wysiwyg thang, i.e. too many writers then thought that they were the re-incarnation of William Blake or Dante Gabriel Rossetti and could both write and do visual design. Or, at least, they thought they could match Jan Tscichold on the typographic design front. Thank god for the Mike Quillmans and Mike Patinos of our world and their visual skill and expertise.

That's a Friday afternoon digression...but then I have plenty of time due to no facebook action. Maybe that's because I was invited to join facebook by my daughter (I told her that my interest was strictly sociological and that I still loved her mother), bless her, or maybe it's because I haven't posted a picture of me or my psychotic dog, or maybe it's because I am a few standard deviations beyond the age of the dating demographic, or maybe it's because I didn't post any java code samples there...

Posted by Jeff Gardiner on October 05, 2007 at 07:39 AM PDT #

Post a Comment:
Comments are closed for this entry.



« July 2016