Each week our WWW stats are up at http://www.fas.org/stats.html. And a log of the Error 404 Missing Links is also available. To figure out who our users are, check out the CyberStrategy Project's Domain Name System guide. Sometimes when our stat.bot is busted, we get stats from our ISP, and we have a little guide to decrypting these cryptic data.
To put these numbers in perspective, Two Commodity Scaleable Servers: a Billion Transactions per Day and the Terra-Server [ A White Paper from the Microsoft Desktop and Business Systems Division], offers some other large numbers of approximately comparable nature.
Estimating the total number of users for a site each week is a bit tricky, since [at least our] log files count unique hosts, which in the case of wetware accesses from corporate LANs or services that use proxy caching [notably AOL] will siginificantly undercount the total number of wetware users of a site. So at least for our stats, these numbers represent an "at least as many as" count of users. RelevantKnowledge provides a Top 25 Web site list ranked by number of Unique Visitors, a statistic similar to cumulative audience in television ratings. For the month of November 1997.
TOP 25 WEB PROPERTIES:
Rank |
Site |
Persons 12+ Cume |
1 |
Yahoo/Four11* |
18,025,000 |
2 |
Netscape* |
13,620,000 |
3 |
Microsoft* |
11,793,000 |
4 |
Excite/Webcrawler* |
11,327,000 |
5 |
AOL.com |
8,764,000 |
6 |
Geocities |
7,830,000 |
7 |
Infoseek |
7,728,000 |
8 |
Lycos* |
5,929,000 |
9 |
MSN* |
5,830,000 |
10 |
CNET* |
5,152,000 |
11 |
AltaVista |
4,727,000 |
12 |
ZDNet* |
4,118,000 |
13 |
CNN* |
3,422,000 |
14 |
Hotmail |
3,041,000 |
15 |
WhoWhere/Angelfire* |
2,780,000 |
16 |
RealNetworks* |
2,541,000 |
17 |
Tripod |
2,377,000 |
18 |
Pathfinder |
2,302,000 |
19 |
EarthLink |
2,257,000 |
20 |
Wired* |
2,236,000 |
21 |
The Weather Channel |
2,201,000 |
22 |
USA Today* |
2,173,000 |
23 |
AT&T* |
2,035,000 |
24 |
Compuserve |
1,934,000 |
25 |
ESPN Sportszone* |
1,870,000 |
This is a measure that aggregates both text and image transfers, which is of interest to FAS since several elements of our site are Imagery Product Libraries
We are also interested in getting a handle on how large sites are, in terms of total number of text files. This sort of information is a bit more difficult to come by, but here are some fragmentary data. The italicized entries are based on AltaVista searches, and, as is well known, this is a notoriously unpredictable service [ie, at one time dang AltaVista only listed about 400 of our over 18,000 pages, and only some 400 of the 13,000 pages @ Spacelink, but 4,000 pages @ Sandia, and a mere 40 @ Sierra Club, so there ....].
We are also interested in getting a handle on how large sites are, in terms of total volume of data on the site. This sort of information is almost impossible to come by, but here are some fragmentary data. As best we can figure, the FAS site has consistently accounted for 0.05% of the entire web for the past two years.
One place to go for comparative stats is Web-Counter, which has a list of the Top 100 sites using their hit counting gizmo. These sites are ranked by hits/day, and one would multiply this number by about six [weekends are only half as busy as weekdays] to get weekly hit counts. This site provides illustrative rather than absolute rankings, since many/most sites don't use this hit counter, and it is clear [I hope] that some of the personal sites of students have artificially inflated their hit counts by starting their odometers with half a million hits [or something to that effect]. But I think that the counts for the internet-oriented sites are probably about right. One other methodological note: these are single-page counts, whereas the other stats discussed here are site-wide counts. Based on our experience base of about three file transfers per user, one would also probably multiply the Web-Counter daily numbers by about three to get total site-wide html file-transfer hits.
Another place to go is the Internet Audit Bureau [IAB], which provides hit counters for a number of sites [you may have noticed a number of sites which carry their little symbol at the bottom of their front page]. They claim to also offer a Heavy Hitters list of the top ten sites for each category of Web sites being tracked by IAB, as well as a Big Hits list of the the top twenty-five most visited sites in each category being tracked by IAB, but as of mid-February neither of these appear to have been implemented. We are not using this service, since their terms and conditions would require us to place an advertisement for some product or service on our front-page.
The Sierra Club numbers are the ones that we are watching closely, since they came online about six months before FAS, and their numbers seemed to approximate the trends we seem to be seeing here at FAS.