Information Technology Services

Web Services

Some Things You Need to Know

Common Terms and Definitions

Web statistics can become quite confusing given the different terminology everyone is using to refer to similar information. This Glossary attempts to define many of the parameters used in Web analysis and what they actually mean.

Hits

Refer to requests

Requests

Requests refer to the number of files requested by each client to the remote server to display the page in question. Also commonly referred to as the number of Hits. An important point to remember about requests is that they do not readily reflect site utilization or the number of users, merely how busy the web server is. This can vary widely between sites since each page is made up of a number of elements such as text, graphics, sound etc. Therefore access to a single page may send 6 separate requests to a web server in order to render the page. A more accurate statistic to use is the number of sessions, since these relate to the number of logins to the site, rather than individual requests. To determine the activity of a particular page it is important to look at the Page report, since these are filtered by the page itself rather than the supplementary graphics and media that it is built from.

Sessions

Sessions are often referred to as visits. This refers to the number of times a user logged into the site over a specified period. This can be done in a number of ways. The most accurate method is to use cookies which track the user as they travel through the site. Essentially, as soon as someone logs in to the site they are assigned a unique id number which is used as a tracking number throughout each session. However, many servers do not support cookies or do not have them switched on. As an alternative method, sessions may be tracked by specifying a session window, which is a period of time that we can designate when a user has logged out if there has been no activity. Therefore when a user logs into the site for the first time we begin tracking them through the site page by page. After there is no activity for a period, say 10 minutes then we assume they have left the site and so increment the session counter. This method assumes that the session window is correct, which may or may not be the case. A session window of 10 minutes is generally regarded as being reasonably accurate.

Sessions are obviously the best method for measuring site activity, since they track each user through the site and can also determine the path each user takes, the time on each page etc. When making decisions about a web site it is more reliable to use sessions as a guide to activity. Note the marked difference in the graphs between requests and sessions.

Visits

Refer to sessions

Report Interpretation

Funnel Web has been specifically designed to assist web site owners to critically appraise the effectiveness of their site. All statistics are derived from the standard and extended access logs from your Web server. Many commercial packages make extraordinary claims about what they can measure and claim to resolve computer locations down to the state level. While this is possible in some cases if the DNS entry in the log file contains state based information, it is more the exception than the rule. As with all statistics be aware of the limitations of such statistics and treat them with caution when using them as a basis for decisions.

Interpretation

With the exponential growth in the number of web sites it is becoming increasingly difficult to attract people to your web site. Statistics provide a new level of insight into your web site and assist in refining and targeting your audience. Careful interpretation of these statistics requires some basic understanding of the limitations of web statistics and what they can and cannot tell you. Firstly, Web statistics provide only the minimum number of visitors to your web site, since many caching systems keep local copies to speed Internet access. But you can be sure there are AT LEAST as many visitors as indicated. You cannot always resolve details on every client access, since many users have either incorrectly setup their domains or simply choose to be unidentified.

Caching and proxies

Most large commercial sites such as America Online, Prodigy, and Compuserve, use large "caches", sometimes referred to as "proxy" servers, on the machines they use to service web requests. That means that once a user looks up one of our pages, it hangs around in memory at that site in case someone else wants to look at it. That way, things run faster - if person "A" looks at our Index page, and two minutes later, person "B" wants to look at it, person "B" gets it right away because it is still in memory from person "A", and the machine doesn't have to get it from us again. What effect does this have? It has the effect of reducing the number of reported "hits" or "requests" we receive from those sites. It's entirely possible that we only register one "hit" for every ten times someone at, for example, Compuserve, looks at the "Index" page.

Users

There is no way to tell exactly how many unique, individual people are using your site unless they are required to login. As noted above, each individual domain represents at least one individual user, but for many users, such as those from AOL, there is simply no way to tell how many different people are represented by that number since users enter your web site via a proxy server. When many requests come from the same machine during a short period, there is no way to tell if it is one user clicking madly or three users browsing at a saner pace. As you can see it is very easy to misinterpret many of these statistics. However you can be sure that they represent a sufficient sample to make informed business decisions about your online activity.

Content last modified: Tuesday, 17-Sep-2002 16:21:39 EDT


	SEARCH: Directory \| Index \| Contact \| Calendar \| News \| Giving