Incorrect correlation using WT.ad and WT.ac in custom reports


Webtrends Analytics 8.x
Webtrends Analytics 9.x


This document explains one common source of data inaccuracy when building a custom report based on WT parameters, dimensions, or measures that are used in one of the built-in reports.


Construct a custom report (and associated dimensions and measures) as follows:

Primary Dimension:
Name: Ad Name
Based on: WT.ad
Collect: Most Recent
Other: Exclude activity without dimension data checked

Primary Measure:
Name: Ad Clickthroughs
Based on: WT.ac
Collect: All hits
Sum across visit: No
Method: Count

Consider a log file with the following data (line numbers included for convenience):

1 WT.ad=Ad1
2 WT.ad=Ad2
3 WT.ac=Ad1

The resulting data in the report will look like the following:

Ad Name

Ad Clickthroughs

WT.ad = Ad1 and Ad2 are in the log, and the clickthrough specifies Ad1. In the Onsite Advertising reports, you will see something substantially different than what has been generated above. However, when using these parameters in custom reports they are treated as any other parameter that might be added into the query string.

Only an Ad Name of WT.ad=Ad2 displays, since that?s the only dimension data that has corresponding measure data. That corresponding measure data is the follow-up WT.ac hit. Note: there is no mechanism to correlate the WT.ac=Ad1 value to the WT.ad=Ad1 parameter like the built-in Onsite Advertising reports do. All the engine is looking for when it decides what dimension value to attribute the clickthrough to is the most recent WT.ad value, in this case, Ad2.

Caution is advised when using a WT parameter in a custom report. Using them to create custom reports is acceptable, but the dimension or measure definition tells the analysis engine what value to pull from the logs and when to do so, not how to use it or how to correlate it with other data. How it is used is based on the report definition. In this way, it is possible to create two reports based on the same parameter and get unrelated results which make no sense when compared. The above report and the built-in Onsite Advertising Clickthrough Rates reports are examples. Additionally, there are multiple, built-in reports that have a specific method of processing data that isn’t obvious in the visible portion of their configuration. The Onsite Advertising and Onsite Search reports are examples. If a report is created using the same dimensions and measures, the results will not display report data as might be expected from the original report. This is due to specific report logic hard-coded into the engine, and it is recommended to use built-in reports when they achieve the intended objective, and use custom reports when built-in reports do not fulfill business needs.