Show TOC

Market Research Hierarchies Locate this document in the navigation structure

 

Market research companies usually deliver the following dimensions in a hierarchical parent-child structure:

  • Product

  • Location (in market research data often referred to as Market)

  • Time

The metadata information of the hierarchies is also included.

Example Example

  • A product hierarchy could have the following levels: Category, subcategory, segment, brand, and product.

  • A market hierarchy could have the following levels: Market, country, region, and Zip code.

  • A time hierarchy could have the following levels: Year, half year, quarter, and month.

End of the example.

Within one data delivery there can be different hierarchies for one dimension.

Example Example

Product hierarchy 1: Product category, segment, brand, and products

Product hierarchy 2: Product category, subcategory, segment, brand, and products.

End of the example.

Market research data is delivered for any level of the provided hierarchies and figures are calculated by the system applying complex calculation rules to the detailed data. Consequently, the delivered key figures cannot be aggregated using the BW Online Analytical Processing technology. Instead the data is shipped as pre-calculated values for different aggregation levels, these are known as pre-aggregated totals.

Pre-aggregation can affect all three dimensions (product, location, and time). The following example shows only one dimension:

This graphic is explained in the accompanying text.

Example of Hierarchy with One Dimension

To ensure correct reporting on different levels in the various hierarchies, an exception aggregation is defined for each key figure and dimension. To report on the pre-aggregated totals, priorities have to be available as reference characteristics for exception aggregation. The calculation of the priorities must be carried out correctly; this happens during data upload in two steps:

  1. During master data upload the system calculates and stores a numeric hierarchy level in the corresponding master data objects Product, Location, and Time.

  2. During transaction data upload the system calculates the priority for each data record.

    Example Example

    Priority = numeric hierarchy level of product * numeric hierarchy level of market (location) * numeric hierarchy level of time

    End of the example.
Hierarchy Metadata

The metadata of the hierarchies of the dimensions product and market are uploaded and persisted in two DataStore objects (DSO):

  • DSO Hierarchy Metadata /DDF/DS71: Contains identifiers and long texts of external hierarchies.

  • DSO Hierarchy Level Metadata: Contains the number of levels for each hierarchy and a long text for the levels.

To distinguish between the dimensions, the InfoObject Hierarchy Dimension /DDF/HDIM is a key field of both DSOs. You can influence the setup of the hierarchy, for example, you can change the descriptions of the hierarchies or hierarchy levels in Customizing or change the sort sequence using a customer-specific BAdI implementation.

Product, Market, and Time Hierarchies

The external hierarchy of market research data is created using the InfoObjects Product (/DDF/PRODREF), and Location (/DDF/LOCREF). To distinguish between hierarchy nodes and leaves, the master data records have a node qualifier for products (/DDF/PQUAL) and one for locations (/DDF/LQUAL). For market research data uploads the value of the qualifiers is set to M. The external product and location hierarchies are converted to a BW parent-child hierarchy using acquisition and propagation DSOs.