How do I filter out excess PDF downloads from WebTrends reports?

Affected Environments

• WebTrends Analysis Suite 7.0x
• WebTrends Log Analyzer 8.X
• WebTrends Log Analyzer Advanced 8.X
• WebTrends Reporting Center 6.X
• WebTrends Enterprise 7.x
• WebTrends Professional 7.x
• WebTrends Small Business 7.x

Cause

The 'Most Downloaded Files' table displays which files are downloaded most often. The table will display the number of Visits and a total number of successful downloads that were achieved by these Visits.

PDF files, act a little differently than most files, however. As a result, these file types appear to have been downloaded with great frequency and with a minimal amount of Visits.

Netscape Scenarios:

When Netscape 4.X or higher is used and a PDF file is clicked while Adobe Acrobat is installed, the file will automatically download to the default temporary directory of the computer. The PDF file will be opened from that directory. This will count as one download and one download session.

If Netscape 4.X or higher is used and a PDF file is clicked while Adobe Acrobat is not installed, the user is prompted to save the file to his or her drive. If the file is saved to the local computer then the scenario will result in the same manner as above.

Internet Explorer Scenarios:

If a user has Internet Explorer (IE) with the inline-version of Adobe Acrobat installed, and the user clicks a PDF file, the file will launch Adobe Acrobat from "within" IE and only download as much as it needs to display the first page and the table of contents, etc. Because of this behavior, it may appear that the one file has been downloaded multiple times since each portion initially loaded is counted as a successful hit to the file. Then, if the user scrolls down to the next page, this next page will be downloaded, and this next page will be counted as an additional hit.

If the user has IE with the inline-version of Adobe Acrobat installed, and right-clicks on a PDF file, saving it to another location, this will count as one hit because the user is downloading the whole file first.

Solution

Workaround for more accurate reports:

When using a Internet Explorer browser, a person who requests a PDF is sent only a single page of the PDF. Each time the user clicks a new page in the PDF, a new page is requested, and that page is sent to them. In the web server log file, each individual request will be recorded as a request for the PDF file. Therefore you may see only a dozen or so sessions/visitors for one or more documents, yet have 200+ downloads on these PDF documents. This is simply because the dozen visitors accessed more than 200 pages of these PDF documents.

There is however a workaround which can be used to more accurately record the number of times the PDF was downloaded. After the first request for a PDF document, subsequent requests show a return code of 206. It is possible within WebTrends to filter out these 206 requests by creating an exclude filter based on return code. Using this exclude filter will result in reports that are more accurate with regard to the number PDF downloads actually requested.