End of Web Design

by Jakob Nielsen on July 23, 2000

Websites must tone down their individual appearance and distinct design in all ways:

  • visual design
  • terminology and labeling
  • interaction design and workflow
  • information architecture

These changes are driven by four different trends that all lead to the same conclusion:

1. Jakob's Law of the Internet User Experience

Users spend most of their time on other sites. This means that users prefer your site to work the same way as all the other sites they already know.

This Law is not even a future trend since it has been ruling the Web for several years. It has long been true that websites do more business the more standardized their design is. Think Yahoo and Amazon. Think "shopping cart" and the silly little icon. Think blue text links.

2. Mobile Internet

Mobility drives small screens (because they are the only ones that can be easily carried) that will often be grayscale (to save battery). Mobile bandwidth will be much more restricted than wired bandwidth. Even though I don't believe in the current generation of WAP phones, I am convinced that mobile Internet will be big once we get better devices - but even these next-generation mobiles will have much smaller screens than PCs. This drives a focus on content and solutions: don't spend screen space on navigation features except for the most necessary ones. With less space for navigation, it becomes more important to stick to standard conventions for where to go and how to explain the options.

3. Network Computing

The network is the user experience that will integrate use of computers and information appliances across locations and devices to form a seamless whole. How can it be seamless if the rules change every time you use a different device? If mobile Internet needs to become simplified, so does wired Internet.

When you deliver a service over multiple devices, users should be able to recognize that it is the same service. Many of the same features should be delivered on each platform, even though some features may be elided or pushed into the background on devices where they make less sense or are harder to deliver with decent usability. These considerations force emphasis on the semantics and not on the representation.

4. Syndicated Content and Services

The days of the unified website are long gone. From about 1993 to 1998, most websites were like Roman military camps: everything inside the barricades was carefully planned and constructed by the residents of the camp. The (fire)wall marked the end of control: everything outside was wilderness and not connected with the site.

This is in contrast to the early years of the Web in 1991 and 1992, where the content on any given server was not connected with anything else on that server to any greater extent than it was connected to the rest of the Web. The Web was a unity and there was no special treatment of pages that belonged to a single site.

Since approximately 1998, it has become more common for websites to rely on syndicated content that flows both in and out of the site. When writing content that can appear on multiple websites, it becomes necessary to restrict the content design to a few mechanisms that will work everywhere, such as headlines, bulleted lists, and highlighted keywords.

Similarly, when a website imports many of its features and content, it typically becomes necessary to conserve resources and expend as little effort as possible on massaging the imports to fit within the site. As long as everything is about the same, it works. Anything too special and you have a conflict.

Application service providers also make it harder for websites to retain overly distinctive design. It is getting to be common for some of the features of a website to reside on other sites that supply certain specialized functionality such as mailing lists, search, conference registration, shopping carts, promotions or coupons, and much more. As users engage outsourced functionality, they would ideally not notice that they have been temporarily moved to a different site for the duration of a certain feature. The feeling should be that of remaining within a single smooth interaction.

Currently, ASPs offer limited means of editing their pages to approximate the appearance of the client site, but it is usually difficult to make outsourced pages feel identical to locally hosted pages unless both have very simple designs.

Expert User Support Requires Sites to Relinquish Control

The last five years, the Web has forced a severe focus on novice users. Basically, all Web users are novices all the time, since you very rarely use any individual website long enough to become an expert user.

Even when some users return often enough to become skilled users, the site still needs to target novices in its design since people will not enter a website if it is not immediately obvious how to use it in a few seconds. Zero learning time or die.

The way to resolve the tension between experienced users' needs for advanced features and first-time visitors' need for extreme simplicity is to move the expert features into the browser or other client software. Two simple examples: the "Back" button and bookmarks. Both work well because they have been removed from the domain of the site and thus work the same everywhere (except for those sites that cluelessly break the standards).

If an expert feature is either standardized across all sites or supported by the client software then it will be available to experienced users without having to be visually apparent in the site design. Thus, it will not cause learning difficulties for novice users. On the contrary, a first-time visitor to a site will be able to use expert features without having to learn the site because he or she can transfer learning from other sites.

What Remains in Web Design

Even as websites become more similar and appearance design becomes more simplified, there will be a large number of design decisions that still need to be made in order to optimize the usability of each individual site.

Most important, each Internet service needs to be based on a task analysis of its specific users and their needs. You can combine standardized user interface elements in many ways, and the better sites will support the way users want to approach the problems.

For example, even if you always call search "Search" and you always have the same way of distinguishing between simple search and advanced search, the question will remain whether advanced search makes sense for any specific site.

Content design will also remain. Each product description is different. Each opinion piece is different. There will always be a need to determine the best approach to describing each unit of information.

Information architecture will partly become standardized. An example that has already happened is the "About the company" area of most corporate websites. All users expect this area to contain subsites about the management, the company history, financial information and investor information, PR and press releases, and employment opportunities. But the way these subsites are structured might differ depending on the characteristics of the specific company. Similarly, there would be many other areas that were related to individual products or services and that would be structured differently on different sites.

See also: Reader comments on this Alertbox (including whether these principles will make the Web dull ) and an essay on the history of print standards.

Share this article: Twitter | LinkedIn | Google+ | Email