Webmaster Bulleted List
Amount Charged for Advertising
Graphs: [Location]
- More Webmasters report not providing advertising on their site (79.25%) than in the Sixth Survey (75.5%). Also, slightly fewer report charging more than $100/week (5.35% Seventh vs. 6.8% Sixth). While this may not seem like a large number of sites, it is more than a 500% increase when compared to the Third years, where only 1% charged this much for advertising.
Intended Server in 6 Months
Graphs: [Location]
- The server that the highest percentage of Webmasters plans to operate in the next 6 months is Apache (35.50%), followed by Netscape (27.06%%), MacHTTP (26.42%), and Microsoft's server (21.84%). Half of the European respondents (51.27%) are planning on using Apache.
Log File Analysis Tools
Graphs: [Location]
- This was a new question to the Sixth Survey. As with the Sixth Survey, most Webmasters reported using a log file analysis package not listed as one of our options (46.26% Seventh vs. 42.94% Sixth), while 28.67% (27.49% Sixth) reporting that they do not use any packages at all. For those Webmasters that do use a package listed, WebStat (13.85%) and WWWstat (13.53) are the most popular. Relatively few users reported using Net.genesis (1.42%), Interse (1.87%), or I/PRO (1.48%). It is our suspicion that many Webmasters write their own custom log file analysis tools.
Mirror
Graphs: [Location]
- The majority of sites do not operate mirrors (87.89% Seventh vs. 85.55% Sixth. This is a very stable characteristic with little growth over the years. In the Third Survey (April 1995) only a small percentage of users mirrored sites (8.55%) then as well (12.11% Seventh). European Webmasters are more likely to operate mirrors (19.62%).
Network Connection Speed
Graphs: [Location]
- Over a third of the Webmasters (38.27%) report that the connection of their server to the Internet is 1 Mb/sec. This reflects an increase from the Sixth Survey where 33.64% reported a 1 Mb/sec connection. The next most common connection speeds are 10 Mb/sec (14.69% Seventh vs. 11.88% Sixth) and 45 Mb/sec (10/69% Seventh vs. 9.14% Sixth). Since the Sixth Survey, there has been considerable movement towards faster connections, reversing the prior trend towards slower servers. Our understanding of this is that, during 1995 and 1996, many commercial sites were going online at significantly slower speeds than those used in educational settings, which had proportionally more HTTP servers in the early days. For example, in the Third Survey, the most common server connection speed was 10 Mb/sec (38.94%) followed by 1 Mb/sec with 12.88% and 56 Kb/sec with 11.08%. In the Second Survey (October 1994), 54% reported connection speeds of 1 Mb/sec! These fast speeds are still typical of the types of connection between colleges and universities and the Internet.
Number of Servers Operating
Graphs: [Location]
- Consistent with the findings of the Sixth Survey, the majority of Webmaster's report operating one server (43.30% Seventh vs. 43.36% Sixth), with 21.46% (23.42% Sixth) operating two servers, 12.89% (13.12% Sixth) operating three servers, and 6.83% (6.56% Sixth) operating more than then servers (a 2% increase over the Fifth Survey). European Webmasters are more likely to operate more servers than US Webmasters.
Operate Intranet
Graphs: [Location]
- Just under half of the Webmasters who responded are operating an Intranet (4.72% Seventh vs. 42.19% Sixth). The number of Webmasters that operate Intranets in Europe is slightly higher: 48.10% Seventh vs. 48.36% Sixth.
Proxy
Graphs: [Location]
- Close to a third (28.93%) of the Webmasters report operating a proxy. The proportion of Webmasters operating proxies continues to grow. (26.08% Sixth vs. 20.0% Fifth vs. 16.70% Fourth). This is especially true for European Webmasters, where 43.67% now operate a proxy compared to 33.60% one year ago in the Fifth Survey.
Sole Webmaster
Graphs: [Location]
- This was a new question to the Sixth Survey. Just under half of the Webmasters share the position with other Webmasters (47.94% Seventh vs. 51.91% Sixth). Europeans are about as likely to share the responsibility of Webmaster (49.09.74%) than US Webmasters (50.91%).
Time Spent Being Webmaster
Graphs: [Location]
- This question was first asked in the Sixth Survey. The majority of Webmasters spend under 25% of their time performing Webmaster duties (59.79% Seventh vs. 60.22% Sixth). Only 9.21% (8.89% Sixth) report spending over 75% of their time as Webmaster. US Webmasters typically spend more of their time as Webmasters than their European counterparts. While the role of Webmaster is a very new position significant portions of these worker's time is spent performing Webmaster tasks, especially in the US.
Which Features are Important
Graphs: [Location]
- Respondents were allowed to mark more than one answer. As with the Sixth Survey, the most important feature in selecting a Web server is performance (83.18% Seventh vs. 80.73% Sixth), followed by ease of administration (73.13% Seventh vs. 72.67% Sixth), price (60.18% Seventh vs. 56.56% Sixth), database integration (53.29% Seventh vs. 53.16% Sixth), remote maintenance capabilities (50.90% Seventh vs. 45.35% Sixth) and secure communications (50.39% Seventh vs. 47.51% Sixth). Virtual hosting increased in importance from being ranked 9th at 28.66% in the Fifth Survey to 7th at 34.34% in the Sixth Survey to 7th at 38.14% in the Seventh Survey. European Webmasters are more concerned about virtual hosting than US Webmasters (48.10% Europe vs. 36.93% US).
Which Server
Graphs: [Location]
- The server used most by the respondents is Apache (34.02%), followed by WebSite/MacHTTP (24.87%), Netscape (23.71%), Other (21.13%), and Microsoft (19.59. This distribution does not mirror other surveys of Web servers, in particular Netcraft's Survey of HTTP Server. Netcraft's survey asks its database of over 750,000 servers what software it is running. Their May 1997 recent results place Apache with 42.36%, Microsoft with 16.31%, and Netscape with 12.14%. There are several reasons why our survey differs from theirs. The most important reason has to do with the non-random sampling and self-selection issues explained in the Seventh Survey discussion of methodology. However, Netcraft's methodology has their own set of troubles, like detecting virtual hosting and load balancing.
Copyright 1997
Georgia Tech Research Corporation
Atlanta, Georgia 30332-0415
ALL RIGHTS RESERVED
Usage RestrictionsFor more information or to submit comments:
send e-mail to www-survey@cc.gatech.edu.GVU's WWW Surveying Team
Graphics, Visualization, & Usability Center
College of Computing
Georgia Institute of Technology
Atlanta, GA 30332-0280