Many companies keep track of their clients’ phone calls and visits to the office, noting by whom they were referred and in what services they are most interested. Surprisingly, some of these same companies pay no attention to the same information that is available for their web sites. Many people probably don’t even realize it, but each time a web page is viewed by anyone, the server records certain information about the request.
Most web hosting providers offer site statistics of some sort. Some offer simple “hit counters,” which are displayed on the page being tracked. A more complete hosting service would include full access to all of a site’s statistics. Being able to access and interpret complete statistics is essential to running a successful web site. What follows is a brief overview of the data that are available, and what they mean.
The web server makes a record in its log for every single page that a visitor requests. Along with the file’s name, size and type, it also records the date and time and several other useful bits of data. The most basic information that can be gleaned from the log is the number of times each page was viewed. By determining which sections of his web site are most popular, the savvy web site owner can focus his time on improving and expanding those sections, capitalizing on their popularity and giving his visitors more of what they want. The most popular pages are the best places to display special offers and important information, and can be used to draw visitors further into the site. It may be surprising to find out that the site’s home page isn’t its most popular page, or even the most common “entry page.” Page popularity statistics also indicate which sections of a site are under-utilized or dormant.
A web server can tell when multiple requests come from the same visitor. This can indicate how many unique visitors a site is getting. Counting unique visitors, rather than page hits, more accurately describes a web site’s popularity, because the count of visitors is hard to artificially inflate. On the other hand, a single visitor reloading a single page over and over again can increase the page hit count significantly! Does this mean that the visitor count will always increase by exactly one for each individual who views a page from the site? Not necessarily. Several visitors who are behind a firewall may appear to be only a single visitor to the server. Tracking unique visits can reveal the habits of individual visitors, though, such as which page they view first, how often they visit the site, and even the entire trail of pages they link through on the site.
There is an intrinsic relationship between page hit counts and unique visitor counts. The ratio of the two indicates how many pages, on average, each visitor views on the site. A high ratio of hits to visitors is generally good, because that means that the average visitor is navigating around the site, viewing several pages. A low ratio, on the other hand, means that the site is probably hard to navigate, or that there isn’t enough compelling information to keep the visitor interested.
Referrals are important to the success of a web site, just as much as they are for any other business. In order for people to find a site, there must be links to it from other pages on the web. These links may be paid advertisements, or they may be the equivalent of “word-of-mouth” referrals. The web server records the address of the page from which a visitor clicks a link to the site.
Since linking is so important on the web, it is important to know which links are actually being used. A web directory listing that costs thousands of dollars per year is worthless if it refers no potential clients!
Search Engine Keywords
In addition to recording referral information, if the referring site is a search engine, the web server will also record the phrase for which the visitor searched. Since search engines are currently the biggest source of referrals for most sites, it has become necessary to “optimize” web pages for search engine keyword relevance. More importantly, analyzing search phrases helps a site’s owner identify its niche on the Internet, and use it to his advantage.
Since the server records the size of each file it sends, aggregating this data gives an estimate of the overall network bandwidth usage for a site. Most web hosting providers have some limit on the allowed bandwidth usage for a site, so it is important to know how much of that allocation is being used. In addition, this can help identify which particular pages or images on a site contribute most to bandwidth usage. Often times, only a few files are responsible for the majority of a site’s bandwidth usage, and those files can generally be made smaller by compression or editing.
Server Response Codes
For every request that is made to a web server, the server will send back a status code along with the contents of the requested page. For a successful request, the status code is 200 (“OK”). This response code is also saved in the log. There are dozens of different status codes, but analyzing a few common ones can provide a wealth of information.
For example, the server will give a 404 (“Not Found”) status and display an error page if a visitor tries to link to a page that doesn’t exist. This may be because someone mistyped the address of the page in the link. Identifying and correcting broken links and other errors ensures that visitors can find what they are looking for.
Historical and Temporal Data
The web server’s log of statistics is saved to a file, and archived periodically. By keeping track of archived data, all of the statistics can be analyzed in a historical and temporal context. For example, it is possible to track usage data during a month-long marketing campaign and compare it to the monthly average, or to another particular month’s campaign, resulting in a simple evaluation of the effectiveness of a marketing strategy.
Historical data can also be used to identify peaks of usage. For example, if a site owner updates her site weekly, and discovers that the site gets more traffic at the beginning of the week, she may decide to publish her updates on the weekend so that most visitors will get the newest information.
While web site statistics cannot be 100% accurate, they still provide a rich overview of a site’s performance and marketability. Identifying the patterns and trends of real usage is a critical element of maintaining a successful web site. The smart web site owner analyzes her sites’ statistics carefully before and after making changes, paying attention to such details as:
Which sections of the site are visited most frequently?
Which pages are most commonly viewed first by an incoming visitor?
During what time periods is the site most active?
How much time does the average visitor spend viewing the site?
When is the site indexed by search engines?
What phrases are being typed into search engines which result in a link to the site?
Where are visitors coming from when they click a link to the site?
Are there errors indicating broken links?
Does the site’s traffic exceed the capabilities of the hosting provider?
By monitoring, understanding, and acting on these data, the web site owner can make informed decisions on routine changes to the site that will engage visitors and influence the rate of conversion from casual browsers to potential clients.