Jakob Nielsen Interview 2 | WebReference

Jakob Nielsen Interview 2

Jakob Nielsen Interview - Part 2

Nielsen Gesturing

WR: There is plenty of emphasis on graphics optimization but little on HTML file size optimization. I've noticed a lot of unnecessary fluff in the HTML of many popular sites (comments, unnecessary spaces/returns). What are some techniques we can use to minimize our HTML?

Nielsen: WebTV has shown the way in this regard: they strip out everything on the server side. It should be quite easy to write a module for Apache to do the same for more traditional Web servers. Considering the importance of download time, it would be great to serve up defluffed HTML.

WR: In order of importance, which interactive elements are most important for increasing Web site traffic? (i.e., links, chat, BBS, popups, etc.)

Nielsen: The only one that is really important is links from other websites. These links can either be based on an affiliate program or they can be purely editorial: if you provide a good resource that other sites feel will benefit their users, then they will link. At least if they have a clue about the nature of the Web. When somebody links to a new site, their users will give them some percentage of the credit for whatever value they find at the new site. Thus, they will be more likely to return to a site that gives them good links.

Of course, you can only attract incoming editorial links by providing good content and good service, and indeed, high-quality content and frequently updated content are my other two recommendations.

WR: For a typical home page, how much would user "click-through" increase by following your usability techniques?

Nielsen: Maybe not at all, if measured in simplistic ways (say, page views). If a site is really easy to use, then users will click the correct button every time and thus need fewer page views to accomplish their goals. It will simply be the *right* page views, as opposed to pages that they may download but don't use.

A better metric would be the number of things users do on the site, whether measured in terms of sales or number of articles they read or something else that actually relates to what users want to do. I would certainly expect any site to be able to double such metrics by employing proper usability. Increases of 400 to 500% would be possible, and sometimes one may see as much as 1,000% improvements. In the long term, I would expect increases of 2,000% or more, but that would require improvements in not just the website itself but also in the browsers and their ability to support human navigation behavior.

WR: What elements of human interaction are not currently available on the Web with our current interface? Which of these interface elements are important to include in the future?

Nielsen: Anything that is highly interactive in a closely coupled loop. Like, for example, the real-time spelling checking in Microsoft Word where it puts squiggles under any typos as they are entered. The equivalent for current Web pages would be a service that noted that you were about to click on a link that would lead to a site that had a low PageRank in Google. If this specific example was all that's needed, then we could have the browser download all the PageRanks for all the hypertext links in a page in parallel with the downloading of the page, but if we needed a few thousand services like this, then it would be impossible to pre-fetch all the potentially required information.

Currently, Web interactions are very similar to the good old IBM 3270 mainframe terminals that we all used (and mostly hated) in the 1970s. As we get faster bandwidth and faster servers, it may be possible to simulate more of the modern GUI-style interfaces that have been demonstrated in university research. But I am skeptical that a "super-Macintosh" style of UI will really be possible over a distributed network. So we may need to evolve a third type of interface that puts most of the speed in the local client and yet gets the appropriate context and content from remote services.

WR: Give us some examples of sites you think are easy to use.

Nielsen: My favorite site was Amazon as of late 1998, before they started to add all kinds of non-book features. A feature like "people who bought this book also liked these books" is a great example of local navigation in the information space constructed by millions of books.

These days, my favorite site is Tomalak's Realm at www.tomalak.org. It is very simple: just a scrolling list of paragraphs, and yet incredibly useful. I particularly like the way juxtaposed links are used to elaborate or comment on the stories. This site is one of the few working examples of a new rhetoric of the Web.

Finally, Google and Epinions are two new sites that both utilize prioritization based on quality ratings. Google proves that it is possible to build a better search engine than the big portals. Epinions proves that user-contributed content doesn't have to devolve into drivel: by empowering readers to rate the quality of the postings, contributors are encouraged to refrain from flaming and anybody who does write a useless posting will plummet to the bottom of the list and be read by nobody.

WR: We've seen some amazing sites with new interfaces using flash, and navigable 3D (http://www.webglide.com, http://www.balthaser.com, http://www.matinee.co.uk/fr_nmd.htm) what do you think of these?

How do you see the Web evolving in the future?

Nielsen: I will answer these two together. I don't think 3D has any kind of major role in the future of Web navigation or general user interfaces. I believe that 3D should be reserved for visualization of 3D objects, such as, for example, showing a scan of the patient's body to the surgeon who is planning where to cut.

Navigation and interaction deal with N-dimensional information spaces that typically become more confusing when they are rendered in 3D than when they are rendered in 2D.

I certainly hope that we will see more good use of animation and multimedia on the Web in the future. Currently, animation is never used to communicate but only to stun the user into submission. And there is no multimedia in the sense I want, which is a true integration between multiple media streams. To me, a video clip is not multimedia: it's unimedia (the medium of television). Video only becomes multimedia if you can interact with it, if it has hypertext links to other media objects, and if it in some ways synchronizes or integrates with other media types.

I see an enhanced emphasis on Web-only content as old media forms like printed newspapers and broadcast television networks die in 8-10 years or so.

And, of course, I see the Web becoming much bigger than it is now, both in terms of size (number of services and users), impact on society and the economy, and in terms of how much and how often we use it. To this day it is still the case that when you think of something you want to do (let's call it X), then it turns out that you can't do X on the Web. Either because X is not available at all or because you can't find it or you can't figure out how to use it. I expect much better search systems in the future that will finally make it possible to find what you want, and I expect that the usability of these many new sites will be at least ten times greater than the usability of current sites.

Finally, I predict that Internet Explorer version 8.0 will be the first good Web browser that actually helps users navigate.

WR: Who do you think would be a better Web designer, Mondrian or Monet?

Nielsen: Monet has a point in his favor because of his willingness to keep painting ever more hay stacks, trying out all possible combinations of weather and lighting. Iterative design is one of the most powerful usability methods, so it is great to have a designer who can produce a lot of variations that you can test with users to discover which one works the best.

The true answer is probably that neither of them would make a good web designer because they were artists which is not what we need for a business website. They wouldn't be good magazine designers either, nor for that matter that great at designing any other utilitarian object.

WR: Thanks for your time.

Note: We'll have a review of Nielsen's new book, Designing Web Usability, later today in our newsletter.

Further Reading

Created: Dec. 13, 1999
Revised: Dec. 14, 1999

URL: http://webreference.com/new/nielsen2.html