Causal data represents constraints, limits, and environmental conditions that can change over time. The condition might be the placement of an asset at a certain location, the price-to-consumer, shelf-space, or one of several other factors that affect your overall performance.
Unlike transactional data, causal data doesn't add up meaningfully over time; instead, UXT calculates causal data one of the following ways, depending on what it represents. Typically, you should not have to worry about how causal data is calculated. Your management defines the method of calculation when it defines the data model, and should communicate the meaning of each causal field to you. In addition, you can place your cursor over the causal data heading to learn more about it.
Causal data from a non-transactional data cube – For this type of data, a starting date and ending date define the time that the condition, or data value, remains unchanged. The data doesn't add up meaningfully over time (for example, you wouldn’t add up shelf space over each day in the month), so UXT weights the data based on the portion of the date range that a condition spans. As an example, consider a customer who allotted 5 feet of shelf space to our brand for the first half of the month, and increased it to 10 feet for the second half of the month. When looking at the entire month, UXT would weight the data to show 7.5 feet of shelf space for the month.
A count of records from a non-transactional data cube – This is like the previous example, but provides a weighted count of non-transactional records for a given date range. For example, suppose that a customer location has two asset placement records for an entire month. The first indicates that an asset was present for the first third of the month and was then removed. The second record indicates that another asset was placed at the location for the last third of the month. UXT would weight the count by time to show that on average two-thirds, or 0.66667, assets were placed at that location.
Counts of unique members (customers, products, channels, etc.) that occur in at least one record during a given date range or pass test conditions are referred to as unique counts. The counts are "unique" because each member is only counted once.
Causal data has its own data category, which is represented by pink. In general, you display causal data the same way you display transactional data.