How to interpret website statistics
It is important to track your website's performance in order to measure the ratio of website visits to sales (or other specific actions, such as subscribing). This will help you monitor the success of your website as part of your overall marketing plan. It is also important to review website statistics in order to find problem areas such as excessive bandwidth utilization and missing pages.
Interpreting website statistics
This article focuses on how to interpret statistics reported by the Webalizer package. Webalizer is a fast, free web server log file analysis program. It produces highly detailed, easily configurable usage reports in HTML format that can be viewed with a web browser. Webalizer typically is installed on servers running the Unix operating system and Apache web server.
Webalizer produces many tables of statistics, but fortunately, as a website owner you only need to review a few items. Below is an easy-to-follow example of where to find relevant information in an actual report, complete with illustrations. It progresses from basic items to more detailed analysis, in the order presented in the Webalizer report.
Viewing webalizer statistics
To view website statistics for your website, follow the instructions provided by your webmaster.
Statistical summary - Usage summary
The first page of a typical Webalizer report shows a summary of activity for the last year. This report contains bar graphs at the top and a table at the bottom. Click on the image to the right to view this image in a new browser window. (Tip: in most current browsers, you can zoom in and out when viewing the image by pressing control + and control -.)
The top right yellow The Usage Summary bar graph shows visits to the website. Visits are the most important statistic. A single visit is someone viewing your website any number of times during a 30 minute period of time. For example, someone browsing ten pages in half an hour would record one visit; someone viewing several pages over one hour would record two visits. (See terms, below. For more information, see Don't Take a Hit on Website Visits - Correct Interpretation of Website Statistics.)
The bar graph shows a peak number of visits early in the year, with a decline during more recent months. The Summary by Month table below the bar graph shows the actual number of visits per month. For example, 32,325 visits occurred in April, 2008. This averaged to 1,074 visits per day.
Do not use the hits statistic, which is very misleading. When a single web page is viewed, a number of hits can be recorded: one for the page, one for the cascading style sheet, and one for each image on the page.
In the Summary by Month table, September, 2008 has a low number of visits because the report was produced during the first week of September. To find the daily average visits, total the visits for all months shown except the top month (September, in this case). Then divide by the number of months and then by 30.25 (the average number of days in a month). In the example, this yields a respectable average of 26,689 visits per month, or 882 per day.
When viewing the main statistical Usage Summary (described above), click on a specific month to view it in detail. This will open a long page of charts and graphs. Fortunately, only a few items need be considered. In our example, we will view statistics for the month of August, 2008. Click on the image to the right to view it in a new browser window.
The first table, Monthly Statistics, shows total, maximum and average statistics for the months. Again, ignore hits and focus on visits, which were 16,054 for the month and an average of 517 per day.
Total KBytes is also important because it measures total bandwidth used during the month in kilobytes (see terms, below). A kilobyte (KB) is 1,024 bytes, and a byte is roughly equivalent to one letter on a web page. In our example, total KBytes is 2,610,180.
Divide total KBytes by 1,024 (or roughly 1,000) to get the approximate megabytes (or MB) - which in our case is 2,610 megabytes. Divide that again by 1,024 to get the approximate gigabytes (or GB) transferred, which is roughly 2.5 gigabytes in our case. This can be compared to the monthly bandwidth allocation for the website - typically 5 to 10 gigabytes per month.
The Hits by response code table indicates whether any problems were encountered. "Code 404 - Not Found" hits track items or pages that were linked to but were not present on the website. Your webmaster should investigate these items.
That's it! You now know how to find the most important items in your website statistics. Continue reading to learn about additional useful information.
The Daily usage statistics are shown as three bar graphs. Once again, the middle yellow visits are most important, showing the number of visits for each day of the month. In our example, daily visits are fairly uniform. The red KBytes bar graph shows the amount of bandwidth used in kilobytes per day. This is the amount of data transferred from your server to people's browsers. In our example, there is one extremely high peak on August 14.
What is causing this? The bottom Daily Statistics table shows visits and KBytes per day. On August 14, 699,828 kilobytes were transferred - approximately 26% of the total monthly bandwidth. But visits on that day were typical, at 525 or approximately 3% of total monthly visits. Thus, a typical number of visitors must have been downloading very large files, such as audio and video.
More advanced analysis
Hits and files can be compared in order to determine a rough approximation of the number of repeat visitors. (See terms, below.) Essentially, a hit is a request for information from a browser and a file is the information returned. However, if a person has recently browsed a web page, the page and its associated images will be stored in their browser's cache, and corresponding files will not be downloaded.
In our example, 11,303 hits occurred on August 1, 2008, resulting in 9,992 files. The difference is 1,311 files that were not downloaded. This is 11.6% of hits. So it is reasonable to assume that approximately 11% of visitors were return visitors.
During that same day, 578 visits resulted in 1,275 pages being viewed, or an average of 2.2 pages per visit. This is notably low, indicating that people most likely found what they were looking for in the first page or two that they browsed, after which they left. It is possible that they did not find what they were looking for, but in that case they probably would have viewed more pages before leaving.
The Hourly usage bar chart shows activity during the day. This chart is useful for obtaining a general impression of when people are visiting the website. In our example, the low point of activity is early morning, with peak activity occurring in the afternoon and evening. Note that the hours shown in the chart reflect the time zone of the server - Eastern time in this case. The Hourly statistics table provides more detailed information for those who have a little too much time on their hands.
The Top URLs table shows the most frequently accessed files and pages. In our example, the home page ("/") ranked first by hits, then the cascading style sheet which is an internal file referenced by every page, then a page called "illegal_alien_numbers.html". Note that this page generated 1,919 hits (not visits), which required 5.48% of total bandwidth to download.
The remaining pages are shown in decreasing order and, with percentages lower than a few percent, are not particularly significant. They do show, however, the pages that people were viewing.
Top URLs by KBytes
The Top URLs by KBytes table shows the most frequently accessed files and pages in order of KBytes transferred. In our example, a video (wmv) file consumed the most bandwidth, then the "illegal_alien_numbers.html" report. Other video and audio (mp3) files consumed a sizeable proportion of the remaining bandwidth used.
Top Entry Pages
The Top Total Entry Pages table shows the entry pages that received the most visits. An entry page is the page that a person sees when they first visit the site (during a half-hour visitation period - see terms, below). In our example, the "illegal_alien_numbers.html" report received nearly 10% of all initial visits, followed by informational pages on Matricula ID cards, drivers licenses, and anchor babies. In other words, people were visiting internal pages as opposed to first going to the home page. This is usually the result of links or search engine results that direct people to specific pages within a website, as opposed to the home page.
The Top Total Exit Pages table shows the pages that people last viewed before they left the website. In other words, this shows where people were when they found what they were looking for or when they got bored and left the website. In our example, the exit pages are the same as the entry pages, meaning that many people went to the website, read one report, and left.
The Top Total Sites table lists the top IPs that accessed this website - essentially the top browser locations that accessed the website. (Think of a site in this context as a location, not as a website.) In our example, the top site accounted for under 3% of website hits, thus indicating that one source is not responsible for most of the website traffic - in other words, there is a diverse range of visitors. The hostname identifies the IP of the website, which can be translated into a domain name with a reverse IP lookup using tools such as DNSTools.com. For example IP "188.8.131.52" resolves to "c-67-185-37-181.hsd1.wa.comcast.net". Your webmaster can do a more detailed lookup.
The Total Sites by KBytes table shows the locations that consumed the most bandwidth. Here, one location consumed 24% of total bandwidth. This quite probably consisted of audio and video downloads.
A referrer is a website that someone was browsing when they followed a link to your website. The Top Referrers table shows the most common websites that people were at when they decided to visit your website. Note that the "direct request" is the most popular in our example - a "direct request" is when a person enters the URL directly in their browser address bar instead of following a link or search engine result to the website.
In our example, the Top Referrers table shows that people frequently found the website through search engines. Note that you can click on the link in the Top Referrers table to actually visit the referring website. Also see Final Note.
Top Search Strings
The Top Search Strings table shows what keyword phrases people were searching for with search engines when they found this website. In our example, the top 3% of the searches were for "anchor babies". The second most popular search was for "matricula".
This table shows what keywords work, but it does not show what keywords did not work -it does not show what keyword phrases people were searching for when they did not reach this website. For example, if people are searching for the phrase "identify theft" and the website is not ranked highly with search engines for that phrase, people would never find the website based on a search for this phrase. For more information on search engine optimization, see How do I get high search engine visibility?
Top User Agents
A user agent is a browser. There are dozens of browser versions and many platforms upon which the browsers run (such as Windows, Mac, and Unix). Popular browsers include Mozilla Firefox, Netscape, Opera, MSN, and MSIE. The Top Total User Agents table shows the most common browsers that were used to access the website. Although a good web designer must routinely account for browser differences and inconsistencies, this table is not particularly useful.
Here are some terms that apply to Webalizer statistics, adapted from the Webalizer website:
You should be aware that while people access websites via their browsers, search engines similarly crawl or "spider" websites. These crawls are relatively infrequent and typically do not inflate website statistics. More information about spider visits can be found by examining website server logs.
Here are links to more information on Webalizer statistics
Contact Elbel Consulting Services for more information on how to track your website's performance.