The usability tests we have conducted during the last year have shown an increasing reluctance among users to accept innovations in Web design. The prevailing attitude is to request designs that are similar to everything else people see on the Web.
When we tested advanced home page concepts we got our fingers slapped hard by the users: I don't have time to learn special conventions for your site as one user said. Other users said, Just give it to us plain and simple, using interaction techniques we already know from other sites.
The Web is establishing expectations for narrative flow and user options and users want pages to fit within these expectations. A major reason for this evolving genre is that users frequently move back and forth between pages on different sites and that the entire corpus of the Web constitutes a single interwoven user experience rather than a set of separate publications that are accessed one at a time the way traditional books and newspapers are. The Web as a whole is the foundation of the user interface and any individual site is nothing but a speck in the Web universe.
I find the current state of Web design to be grotesquely inadequate in supporting navigation and other user needs, so I am somewhat depressed about this growing user conservatism. It would be unfortunate to freeze Web design while it is still in its infancy. However, I have always operated under the dictum that the user is always right, so when users demand increased compliance with conventions, we have to follow suit in any individual design. After all, the Web is greater than any one of us, and we can't expect our site to be used if we reject the Web and build our own thing.
On the other hand, I am not willing to stop progress at this early stage. My solution is to call for modest incremental advances in Web design, with new sites mainly following conventions while pushing a small number of new ideas. Some of these new ideas will be so successful that they will emerge as conventions in their own right. Also, there is much potential for improving the Web without changing its surface appearance. For example, better writing has been shown to more than double the measured usability of a site. Also, better search algorithms, using ideas like synonym searches, spelling checks, relevance ratings, and reputation management can improve the quality of searches on the server.
Slower Uptake of Software Upgrades
The figure shows how many Netscape users have been using its various versions over time according to data from Interse and AdKnowledge. For all subsequent releases after Version 1, the figure shows a fairly similar picture: a slow increase in use of the new version combined with a corresponding drop in use of the previous version(s). The slope of the uptake curve for a new release corresponds to the users' upgrade speed. For versions 2 and 3, the upgrade speed was 2% per week. In other words, every week, about 2% of the users switched from previous releases to the new release. Moving most of the users to the new version would take a year (about 50 weeks at about 2% per week).
For version 4, the upgrade speed has dropped to about 1% per week. With this reduced upgrade frequency, it now seems to take about two years to move most of the users to a new browser version. This reluctance to upgrade is probably due to four factors:
- The user base has moved beyond the net-nerds who are interested in the Internet for its own sake. In 1993 and 1994, I would go to the Mosaic FTP site almost weekly to see if they had a new upgrade available, and in 1995, rumors of impending Netscape beta releases were treated as Usenet gold. In contrast, most current users are interested in what the Web can do for them. They care about content; not about technology or software. Therefore, they are not as motivated to upgrade.
- Many new users do not know how to upgrade their browser, so they stay with whatever version is installed on their machine. Early Internet users were skilled in the arcane arts of software downloads, installation, changing default preferences, IP numbers, proxy settings, and much more. You had to be in order to be on the net. Now, it has gotten easier to get online, so many users are less skilled in the guts of the technology.
- Recent browser upgrades have been less compelling than previous ones in terms of added features or improved usability. The earliest browsers were very primitive and the relative improvements from one release to the next were fairly large; leading to greater benefits from upgrading in times past.
- The size of the download in megabytes has grown faster than the available bandwidth, thus making it more costly to upgrade.
Even though the data in the figure relates to a specific browser, it is likely that there are similar slow-downs in upgrade speeds for other browsers and for plug-ins. Because of the slower upgrades of Web software, websites will have a larger number of users who continue to use older versions. This again increases the need to design according to established Web Consortium standards.
Conservative Client Design; Aggressive Server Design
Web users are conservative: they don't want inconsistent site designs or fancy pages filled with graphic gimmicks and animations. And they frequently don't have the latest client software available. As a result, Web designers have to be conservative in what they show to users: page design must be conservative and minimalist. The road to better Web design is to move complexity and advanced design to the server and provide more intelligently processed services to the users.
Update added April 1999: newer statistics 13 months later confirm and extend this analysis.