The Webtrends base license key (authorization key) allows for two concurrent activations of the product. One is intended for production, the other for development. The development server can be used for test analyses without affecting the licensed page view count of the production server. To estimate annual page view usage, perform the steps below:
1. Install Webtrends on the development server as a stand-alone installation (stand-alone meaning that the installation does not reference the system database on the production server, but installs its own).
2. Gather one week's worth of log file data from each web server that will be undergoing regular analysis and place the collection of log files from each server in a separate folder in the data source location on the development server.
3. In the Webtrends user interface, navigate to "Administration > Application Settings > Data Sources" and create a new data source for each folder of log files, prefixing each with a number such as 01-, 02-, etc., (to mark increments), then use a naming convention to indicate from which of the web servers the log files originated.
4. Under "Web Analysis > Reports & Profiles", create a profile using the first of the data sources created in step 3, and name it after the data source for easy identification. When the profile is saved, analysis will begin (by default).
5. After analysis has completed, open a command prompt and navigate to the following location in the Program Files folder:
6. Type the following command, then hit Enter:
wtlicman -s > license.txt
This will create a text file in the \lib folder named license.txt containing verbose licensing information, including a running total of licensed page views used by the installation.
7. Copy license.txt to a different location and name it after the profile that just completed analysis.
8. Repeat steps 4-7 above until each profile has analyzed the contents of its corresponding data source and a version of license.txt has been generated.
9. Review the first of the text files generated. At the bottom, under "Standard pageviews collected:" and "SDC pageviews collected:" a count displays for the anniversary year beginning the month of installation through the preceding month of the following year. "Standard pageviews collected" refers to log files generated by Apache, IIS, and other logging solutions. "SDC pageviews" are from log files created on Webtrends' SmartSource Data Collector. Whichever of the two values for a given period is higher will be the one that determines licensed page view usage. Multiply this value by 52 for each week of the year. The resulting value will be a projection of the web server's page view usage throughout the course of the year, assuming the contents of the week analyzed are representative of weekly traffic on average.
10. Review the other text files in sequence, subtracting the previous total from the total in the file currently under review, then multiply the value by 52. This will be the estimated page view usage for each server.
If you find an anomaly like a profile with an unexpectedly high volume of licensed page views, and the Pages report displays a much lower figure, review the Accessed File Types report under "Site Design > Pages and Files". This report may display the activity the Pages report does not, as the Pages report is based on known file types. File types that are unknown, and not among the exclusions on the NonPageViewExtension line in license.txt, will still be counted toward the licensed page view count.
If the web site(s) producing the log files come under increased load for certain events or on a seasonal basis, it may be helpful to perform a test similar to the above using the log files from the previous year with a larger reporting period (such as a month or quarter) to make a more accurate projection. An example of this would be web sites that are busier during the Christmas season. If a projection was made that did not account for the increased traffic in this period, it may mean licensed page views are exhausted far earlier than anticipated. If this scenario applies to the sites in question, it is recommended both "low season" and "high season" tests be made and a value extrapolated from summing the ranges to which they apply.