Website Traffic Analysis: Is Weblog Analysis Adequate?
Summary : The Internet bubble may have burst, but Internet marketing is stronger than ever. Paid search engine advertisement is growing at a stratospheric rate of 25% per quarter (not per year). Internet sales are growing at a rate much faster than overall retail sales. The Amazon stock has nearly tripled in a year.
The Internet bubble may have burst, but Internet marketing is stronger than ever. Paid search engine advertisement is growing at a stratospheric rate of 25% per quarter (not per year). Internet sales are growing at a rate much faster than overall retail sales. The Amazon stock has nearly tripled in a year.
What is happening is that all organizations, even small ones, are realizing that Internet hype may have been overdone but the importance of Internet marketing is growing stronger every day. Internet marketing is all about bringing traffic to your web site, and then converting this traffic into achieving your strategic goals, be they selling goods, building awareness or delivering information.
In this context, understanding traffic to your website gains critical importance. Simple metrics such as counters or page views are totally inadequate in the marketing context. Marketing managers and small business owners need to know who is visiting, from where and when, why they are visiting and what they are doing when on the website. These are the 5Ws of Internet marketing.
Currently this type of analysis is being conducted in two ways: Weblog analysis and Tracking code analysis. The web logs are provided by Internet service providers, and there are both online and offline tools to conduct the analysis. Tracking code analyses are provided online by several providers such as Hit Box and Web Site Traffic Report.
Although both types of analyses are supposed to count the same thing (visitor actions on the website), there are technical differences that are subtle in scope, but significant in impact. Tracking code systems are generally more direct and more accurate. Further, tracking code systems are relatively so inexpensive (usually starting at $10/month) that every organization should investigate the value of their use. The cost of using inaccurate statistics far exceeds the minor cost of tracking code systems.
The table on the following page highlights the main technical differences between the two systems and their impact on the outcome.
| Feature | Weblog System | Tracking Code System | What does it mean? |
| HTML Frames: Often each webpage may have two to four or even more frames | Often counts each frame as a page view. | Page is only counted once. | Serious over count with weblog analysis. A serious discrepancy. |
| Cached pages | Many ISPs store many pages on proxy servers to improve performance. These are mostly not counted by weblog systems. | Properly counts each as a page view. | Undercount by weblog analysis. The seriousness of the undercount depends on the use of proxy servers. |
| IP address pools | Many ISP have a pool of IP addresses that change over time and even during a session. Weblog systems are based on IP address and thus may over count users. | Uses an internal session cache that does not rely on IP address. | over count by weblog analysis can be significant. |
| False page views | If the users quickly skip over pages even before the page is fully loaded, the weblog system count the skipped over page as a page view. | The page view is not counted until the tracking code (usually at the bottom of the page) is loaded. | over countby weblog systems can be significant. |
| Artificial traffic | Scores of search engines spider the web every day. Any hit by the spiders is counted as a page view. | The spidering is usually too quick to be recognized by the tracking code systems. | As the number of search engines grows, the over count by weblog systems can be serious. |
| Placement of tracking code | No impact. | In some very slow systems, the tracking code at the end of the page may not be recognized before page change even when most of the page has been read. | This is a relatively rare problem and is even of less significance with faster connections. Tracking code systems may yield a minor undercount, if any. |
The analysis above clearly shows that tracking code systems provide better data that is the basis of all analysis.
All tracking code systems are not alike. Many produce scores of tables and charts, but the critical insights are lost in charts and graphs. We recommend Web Site Traffic Report (www.websitetrafficreport.com) where the key data is intuitively organized in a single report that is automatically emailed to the users every day while the users still have the options to go online at any time and conduct additional analyses.
No comments:
Post a Comment