Webtrends Analytics 9 Adminstration Guide, November 2012
To create meaningful reports from the enormous number of combinations of dimensions and measures allowed, think about what kinds of dimensions and measures work together meaningfully. Combining a higher-variability dimension with a lower-variability measure can lead to inaccurate data.
Dimensions and measures can be classified as hit-based, visit-based, or visitor-based.
|Hit-Based||Changes from hit to hit within a single visit. With this type of dimension, y you can only use hit-based measures.||Content groups, products, and campaigns are all hit-based dimensions because it is possible for each a visitor to click in a different one for each hit. Hits, page views, and clickthroughs are all hit-based measures.|
|Visit-Based||Does not change or occur multiple times within a single visit. With this type of dimension, you can use visit-based and hit-based measures.||Each visit typically has only one referring site and one geographical location such as a country. Visits, visit duration, visit depth are all visit-based measures.|
|Visitor-Based||Does not change for an individual visitor, even across visits. With this type of dimension, you can safely use visitor-based, visit-based, and hit-based measures.||Each visitor has only one initial campaign or initial search engine. Visitor-based measures include unique visitors and unique buyers.|
Measuring a visit- or hit-based dimension using a visitor-level measure such as unique visitors produces inaccuracy in proportion to the percentage of repeat visits within the chosen time frame. For example, if you create a report using content groups (hit-based) as the dimension and unique visitors (visitor-based) as the measure, your data becomes increasingly inaccurate over time as the same visitor returns to the site and views different content groups.
If you want to report on visits rather than individual hits, you can choose to sum a measure across visits. Summing is a measure-specific setting that determines whether Webtrends adds all the hits in a single visit and treats them as one entity. Summing across the visit assigns the sum of that measure for all page views during the visit to the corresponding dimension. Because this setting is only significant when Webtrends counts a measure more than once per visit, it is only available when you click All Hits or Hits That Match Specified URL in the Analyitics 9 When to Measure dialog ( , click a measure, and then click ).
The following example shows a dimension that does not vary during a visit. The Revenue (sum) and Revenue (no sum) measure values are the same for each dimension item, but note the difference when the measures are calculated as averages. For Average Revenue (no sum) the average revenue per order (not per visit) for Missouri is $25 ($10+$20+$30+40)/4=$25). With summing across the visit enabled, however, each order event is assigned the sum of the purchase values for the visit, and the Average Revenue (sum across visit) is calculated as ($30+$30+$70+$70)/4=$50.