The Washington Post
Navigation Bar
Navigation Bar

Enter symbols
separated by a space:

Look Up Symbols

Made Possible by:

Related Items
 On Our Site
  • Most recent .com column
  • TechThursday
  • E-mail

    Hosted by Leslie Walker
    Washington Post Columnist
    Thursday, June 17, 1999 at 1 p.m.
    Leslie Walker
    ".com" columnist Leslie Walker

    Welcome to ".com - Live," a real-time, moderated discussion with the people who are shaping the business strategies in the era of electronic commerce. My guest this week will be former Sun Microsystems engineer Jakob Nielsen.

    Nielsen is on a crusade to make the World Wide Web a friendlier place. Using his Ph.D. in "user interface design" -- the face that computers present to people -- Nielsen advises businesses on how to make their Web sites easier to navigate.

    Jakob Nielsen
    Jakob Nielsen

    Nielsen believes the vast majority of Web sites are so badly designed that customers can't find what they're looking for or accomplish what they are going online to do.

    We'll be talking about the top mistakes of Web design, how the Internet architecture is evolving, and how the structure and design of the Web affects the way people use it.

    Transcript Follows

    Leslie Walker: Hello everyone and welcome to our Web talk show. We're delighted to have Jakob Nielsen here today to help us understand the chaos that is the Internet. Jakob left his post as a distinguished engineer at Sun Microsystems last summer to become a Web design consultant, advising corporate America on how to drive the bumpy roads of cyberspace. Jakob has been studying information design since before the personal computer was born in the early 1980s. He holds strong views, like his belief that most Fortune 500 companies will either go out of business or shrink over the next decade because they can't their customers' needs online. He'll be talking about those views and more today, so send in your questions now!

    Leslie Walker: So, Jakob, let's start with Web architecture. Why is it so hard to navigate and find stuff online? Is it because of how the network developed historically, or do you think the chaos was/is avoidable?

    Jakob Nielsen: There are two reasons: the first, and the easiest to fix, is that most companies use "ego-centered design" instead of "user-centered design". In other words, if the designer and/or the VP of marketing thinks that the site looks good, then it *is* good. Not only does this lead to bloated pages that take forever to download (unless you happen to use a machine with a direct connection to the server), but it also leads to information architecture hat mirrors the company's internal structure and not the way users think about their tasks. If the structure is bad, then of course it's difficult to find anything. The solution is reasonably easy: just integrate usability engineering into every part of the Web project and make all decisions based on how they impact users in the tests.

    The second reason is more fundamental: The Web (and the Internet in general) currently does not have software that helps people find things or navigate the *huge* information space. Search engines are a joke that worked well back when the Web had 10,000 sites but have not scaled to the current 6 million sites. Browsers should really be called "page viewers" since they don't help on the *browsing* task except for very primitive tools like bookmarks and the BACK button. We have had these same tools since Mosaic in 1993 when the Web has 1,000 sites. They worked well then, but we need much more now.

    Finally, the architecture is based on the assumption that all data is created equal. But it's *not*. Some pages are good and some are worthless. The new search engine Google uses quality ratings to point you to the best stuff first, but thatís a rare exception. We need more tools that guide users to the best sites and that help them understand what e-commerce sites are credible and which are not worth your dollars.

    Leslie Walker: You recently studied 20 high-traffic Web sites to see how easy to use they were. You found more than 80 percent had page download times of 19 seconds or more--way more than the ideal minimum you recommend of eight seconds.

    Why is that such a big deal? And isn't it a lot better than it was, say, a year ago?

    Jakob Nielsen: I think it *is* getting better: a few years ago truly bloated pages with one huge image were common and download times often ballooned to the 30-40 second range. So many sites have recognized that looks are not everything.

    But itís not getting better fast enough. 8 seconds is about the delay tolerable before people start leaving a site. After 10 seconds, the human attention span stops and you canít keep focused on using the site. But that does not mean that all is peachy if only pages download in, say, 7 seconds. The true finding from human factors research is that we need sub-second response times for the Web to feel pleasant to use and for users to browse freely. 19 seconds is 1800% too slow. Not a small deal: a big deal. Users who get DSL or cable modems say that the one most striking observation is how *different* the Web feels when pages come as you click rather than involving a long wait.

    Vienna VA: Hi Jakob! I read that ZDNet's Jesse Berst called you the smartest man on the Web.

    So can you tell me how BIG is the Web today, how BIG is it going to be in a few more years, and is there any chance it will collapse under its own weight, either literally or figuratively?

    Jakob Nielsen: Right now, the Web has 6,177,453 sites according to the Netcraft survey (

    With this growth rate, it should hit 10 million early next year and 100 million in 2002.

    Usability definitely will suffer under this immense growth unless we do get better tools in areas like search and navigation.

    Stockholm, Sweden: What makes users trust the information carried out as text on a website? Is a user-friendly website more trustworthy than a non-userfriendly? Is there a connection?

    Jakob Nielsen: Yes, ease of use definitely adds to credibility. If something is difficult to use or users feel lost all the time, they get the feeling of being poorly treated. Also, users feel "if these guys can't even design a website, then can I trust them with my money? Or if I order something, will the product be messy as well?"

    Another point is to be less promotional in the language on a site. No "marketese" like you often see in sales brochures. Users equate such language with the hard sell and immediately start distrusting the site: "What are they trying to pull on me?"

    Seattle,WA: What are the major steps you would go through from scratch in an evaluation of a Web site? Should it be just "transparent" or more like an
    "engineered experience"?

    Jakob Nielsen: The first step is always a user test of the old site (assuming that it's a redesign project and not a brand new site). The second step is a user test of competing sites (this can also be done for a brand new site since there are *always* other sites in the world that do somewhat similar things).

    Whether running a user test or doing an expert evaluation, the first things to look at are always whether people can complete common tasks easily and quickly. Later on, you can look at whether users can do complex tasks or use special features. But if the most basic things are not easy, then nobody will hang around long enough to even get into the advanced features.

    I am more in favor of transparency than engineered experience because I want to support users in doing the things *they* want to do. After all, that's why they are on the web in the first place.

    Chicago IL: Hello Mr. Nielsen. What tips can you offer us to speed up our Web navigation? And are there one or two search engines you consider the best?

    Jakob Nielsen: There is no doubt that the best search engine on the web these days is Google ( Disclosure: I do sit on their advisory board.

    I also like InfoSeek (now called Go) and AltaVista. Excite also has a very good feature that may be unique in its ability to search current news. Very valuable for finding coverage of current events or new products or anything else where you can't wait a month for the regular search engines to update their index.

    The truth is that there is no perfect search engine today and you need to use a combination: get to know several services and use each one for what it's good for. But Google is a good starting point (except when looking for current events): you will almost always find what you want in the top 2-3 results it returns.

    Houston, Texas: Jakob,
    what is your take on Seth Godin's idea of permission marketing? Should websites just be a vehicle to register customers and not much more, as Godin suggests? Or is content still critical for attracting regular traffic? Also is this why you don't ask people to register on your site-


    Jakob Nielsen: Permission marketing is much better than the prevalent approach on the Web with advertising and spam. I think ads are doomed on the Web since users basically ignore them. But when a user *wants* something, then you can sell them!

    It is important to be sure that you donít just get *permission* from the user but have an active *desire* for communication. If people grudgingly give you their email address in order to get some benefit, then they are not really looking forward to your messages which they will treat like so much spam and delete immediately. But if the users have *asked* to be contacted because they *want* the info, then you are in marketing heaven. The best current example is Amazon.comís announcements of new books by your favorite authors:
    (*) Itís pure opt-in
    (*) People *want* to know when a new book is out by their favorite author
    (*) Itís a service to the customer to allow them to buy the book with a single click

    Gaithersburg, MD: In general, what do you think of the navigability of government websites compared to those of the private sector?

    Leslie Walker: Good question--I'm also curious, are there any government Web sites that you consider particularly well designed?

    Jakob Nielsen: I think they suffer even more from ego-driven design than many commercial sites. And the government doesn't go out of business because people don't come to its sites, so there is less of a motivation to get things fixed.

    There is also the problem with internally-oriented language and internally-oriented structure: the government has it's own language which they all think is natural. Well, it's not. No normal person understands the acronyms or which departments do what.

    A final problem is the desire of each department and division and office to have its own design and own rules. So every second click, the user is faced with a completely different set of navigation principles, features, and layout. Such inconsistency truly hurts usability.

    I don't like the design, but it's nice that the IRS has all the tax forms available for download in a standardized format. But the interface is definitely very internally-oriented since you have to know their stupid form numbers to get anything. It's also very clumsy to navigate.

    Some of the smaller sites are better: For example, the U.S. Mint sells collectible coin sets online. That's pretty much up to the standards you expect from a good e-commerce site.

    NC: Why is it so hard to convince people -note: managers- that usability matters? You mentioned the "ego-driven" concept of design. How does one who is concerned about usability convince those who hold the cash -usually not short on ego- that such work will not cut it? This isn't "Field of Dreams". They are not going to come just because you build it. Why is that such a hard concept for many money-mongers to get?

    Jakob Nielsen: In the long term, sites that are not user-centered will die the death of the million mouse clicks as users go straight for the BACK button.

    This is why sites often get the usability religion on their third iteration. No matter what they do in terms of going to a different agency with a different glamorous design, they still don't get any users or any business. So they finally recognize the need for ease of use.

    The question remains, are companies doomed to always have at least two bad versions of their site before they can design something useful? I think yes, because the Web is so different from anything management has seen before that all their instincts are wrong and lead them toward disaster.

    It's hard to convince managers that usability matters because they take one look at the site and say "but that looks easy enough: I can figure it out". Of course they can: they already have all the right mental models. You can't wipe your brain and pretend that you don't know what you do know and assess whether a new customer would be able to find anything. (Unless, of course, you are a usability expert, but it takes years to get this skill.)

    Leslie Walker: You say banner advertising on the Web is doomed. As we all know, a gazillion new Web sites are built on the promise of ad revenue. Do you envision WEb advertising reinventing itself, alternate revenue streams emerging-- or a whole lot of Web sites crashing and burning?

    Jakob Nielsen: Crash and burn time.

    Web advertising is doomed, not because of poor banners, but because it is a fundamental mis-match with the way people use the Web. Users go online to get things done, to find something, to read something. Whatever: the key fact is that they are goal-driven. They are not going to take the time to click elsewhere.

    Even the best creative will not work because users have been burned in the past by bad advertisers. Once you know that it is usually not worth paying attention to Web ads, you start ignoring them. Selective attention is very powerful.

    The solution is not ever-more intrusive forms of ads. We can have an arms race between more bouncing and jumping ads versus users getting more thick-skinned and better-able to ignore the ads. Or we can accept the data that shows that Web advertising doesn't work and start on other business models. My favorite is micropayments.

    Fairfax, Virginia: Can you discuss any guidelines about color, layout and metaphor that make a website user-friendly from an international viewpoint.

    Jakob Nielsen: The biggest is that the layout needs to be mirrored in countries that read right-to-left (e.g., Japan, Israel, all Arab countries).

    It is true that there are differences in color preferences between countries, but the biggest issue is still to match users' local needs and local culture. Even a simple think like disclosing international shipping rates if you are an e-commerce site. Or having electronic equipment for sale that works with both 110 Volts and 220 Volts.

    And test for international usability. If native speakers of your language have difficulties using the Web, just think how bad it is for people who speak a different language to access your site. even when the language is the same, there are still differences in behavior and conventions. So don't think that a U.S. site will be perfect in the U.K. without a usability test with British users.

    Seattle: How much of a performance improvement do you expect when
    you redesign a major website,
    and how do you measure the

    Jakob Nielsen: At least 100%. It is very common for usability to be doubled with even the smallest improvements in usability: fix the language, write headlines that make sense, tweak menu items, and such.

    Bigger changes could be to get an information architecture that matches users' needs and a page design that downloads *fast* and makes the most important elements the most prominent while keeping a consistent placement of all other options. Such bigger redesigns typically result in between 300% and 1000% improvements in usability.

    Usability can be measures most simply as the time needed for users to perform a standard task. For example, start at the home page and buy an airline ticket to London.

    You can also measure the percentage of users who are capable of using the site at all (typically, less than half the users will be able to do something if you take them to a home page and let them loose).

    The ultimate measure, of course, is increased sales or other business measures. Again, several hundred percent is typical.

    Leslie Walker: Micropayments is an interesting idea, but there are so many competing standards for how Web sites might collect small payments for slices of their content. Is it realistic to think the decentralized, ungovernable Internet can develop a common set of micropayment standards? Do you see the seeds of this sprouting somewhere yet?

    Jakob Nielsen: We do need a standard. The main problem right now is too many different micropayment systems. It's like needing a wallet full of ten different currencies just to shop at a single mall.

    I think we will see one or possibly two standards in a year or two (possibly two standards: one for truly small payments and another for larger amounts). We may need a coalition of technology vendors (Microsoft etc.) and content vendors (Disney etc.) to push this through. The payment vendors may not be able to do it alone.

    But without micropayments the web will never reach its full potential. There are so many new services and business models that become possible once we can get paid for giving service to customers. Right now, the Web is in a state similar to the barter economy: I give you a pig (tolerate a blinking banner) and you give me hundred bananas (good articles). Well, for a country to grow its economy, it needs a money economy. The Web does as well.

    San Francisco, CA: Do you think there is a chance the Web search engines will become so full of irrelevant and outdated links that people will stop using them?

    Jakob Nielsen: Yes. Unless you reliably find useful results when going to a search engine, it's not worth coming back.

    This is why I am so optimistic about the prospects for better search engines. The old ones will simply not be able to handle a Web with 100,000,000 sites unless they drastically improve their technology. With more stuff, better quality control becomes ever-more necessary.

    Leslie Walker: Web audience measurement firms are churning out statistics on how many people visit every site each month, how many pages they look at, how long they spend on each page. In your opinion, what measurement data is most important and why?

    Jakob Nielsen: Many of these measurements are irrelevant. Take "reach" which is reported so widely. Why should it matter how many users have been to a site once (so-called "unique users") but have left it in disgust due to bad design? On the contrary, counting unique users rewards a site for bad design and is a very misleading metric.

    If you believe in Web advertising (I don't), then reach is still irrelevant. It doesn't matter how many prospects see *other* pages on a site than the ones where you ad runs.

    Two better measures are:

    (*) Loyal users who use a site repeatedly

    (*) Total time spent on site

    (Total time risks being misleading as well since you can increase the measure temporarily by making the site more difficult to use so that people will need to look at more pages to complete their goal. Of course, users will be less likely to return another time, so it's not good in the long run.)

    The ultimate measure is loyal users: How many people see this site as a part of their Web lifestyle? Do they use it often (very loyal users) or less often?

    Washington, DC: The rule of thumb for Web design and building has been to put the bulk of your money in to supporting maintenance issues rather than initial design. Is that rule of thumb still appropriate today and what sort of ratio should we be considering?

    Jakob Nielsen: My advice has always been to budget for annual maintenance costs of 50% and 100% of the money spent to develop the site in the first place.

    Keeping a site fresh and updated is essential to keep loyal users coming back (see previous question).

    A recent survey by the Association of National Advertisers (ANA) showed that their members had maintenance budgets of 72% of the cost to develop a site in the first place.

    Bethesda MD: I am overwhelmed with all the information I get every day. Email, Web pages, faxes. Are there any new filtering programs that you recommend, especially for email? I read about all these fancy filtering capabilities, but who has time to buy and test a bunch of new software?

    Jakob Nielsen: You are not alone! It's getting to the point where simply going through your email means a day's work.

    Add voice mail, faxes, pagers, and so on, and we will never get time to think new thoughts or get any real accomplishments done.

    Unfortunately, current email filters are not very good. They do allow you simple things like putting all messages from certain mailing lists into a special folder so that you can look at them all at once when you have time. And spam filtering is getting better as well, allowing you to kill many "get rich quick" schemes.

    But much more progress is needed. The underlying ideology of email for the last 30 years has been "the mail shall go through". You know, like the Post Office slogan; neither snow, nor rain, nor gloom of night shall stray these couriers from their appointed rounds.

    In the future, we need an email architecture that is aimed at optimizing the recipient's time and productivity. For example, when a sender shoots off a message to 10,000 employees in a company, he or she should get a dialog box saying "You are about to spend $5,000 of the company's money."

    And I also believe in pay-for-delivery schemes (electronic postage stamps) where you have to pay me to put something in my in-box. Maybe 5 cents per message or something. That would kill spam quickly, for sure.

    Leslie Walker: WE are just about finished here, folks!

    Leslie Walker: One final question for the Web's premiere usability guru: Can you peer into your crystal ball and tell us what kind of Internet you envision in, say, five years? What macro trends do you see shaping it?

    Jakob Nielsen: Bigger, bigger, bigger.


    But bigger doesn't just mean more of the same. It means many more services that integrate much more with out daily lives.

    Macro trends:

    More international users: of the 1,000,000,000 users on the Web in ten years, only 200,000,000 will be in the United States. 80% will be elsewhere.

    More mobile use, stating this year. The Web will no longer be an exception, it will be with you at all times. This will change use drastically.

    The Web becomes the default way to do business, whether business-to-business or how consumers shop.

    I think the Web will become like electricity: an integrated part of your life which you use all the time without even thinking much about it.

    And then there will be new services that will be wonderful. Count on it! (At least if we get micropayments so that the service providers have a way of getting paid)

    Jakob Nielsen: Thanks, folks. Great questions. Sorry I couldn't get to them all. There is no doubt that the Web has a lot of momentum these days,

    Remember: If you make usability the driving force behind your website, then not only will your users love you,but that's also the way to make them come back and to make money. It's a true win-win.

    Leslie Walker: That is all we have time for today, folks. Sorry we couldn't get to every question you sent in, but I will forward the unanswered ones to Jakob so he can see what was on everyone's minds. Thanks to the former distinguished engineer from Sun for taking time to explain his view of the Web. We hope to have him back again soon! Hope to see everyone again in two weeks.

    © Copyright 1999 The Washington Post Company

    Back to the top

    Navigation Bar
    Navigation Bar