Let’s talk about Attributes and Hierarchy
Two of the most important concepts in marketing analytics are attributes and hierarchy, both of which drive the structure of the data and how it is interpreted and turned into decisions. If either the attributes or the hierarchy aren't well-defined, then the analysis will be disjointed, and the insights that are produced from the analysis will not provide enough information at a strategic level.
Attributes are simply the characteristics of the data that are used to describe, group, or differentiate the data from another data point. In terms of attributes in a marketing sense, they include areas such as type of product (brand), size of the product, price category, channel, geography, promotion type, or customer type. Attributes answer the question “what” is creating results. For example, simply being able to see that there is revenue is not very informative. However, seeing revenue by channel, by region, and by customer type gives insight into how to drive the results. Proper data analysis will depend on selecting the right attributes to achieve the business objectives; therefore, tracking everything available at once does not equate to gaining better insight.
Hierarchy provides structure to the attributes by grouping them in a way that matches how decisions are made. For example, a typical product hierarchy would go from the product category level to the brand level to the SKU level, while a geographical hierarchy typically goes from country to state to city. Hierarchies answer the question, “Where do I begin my analysis?” They provide support for analysts to zoom out to look for strategic trends as well as zoom in to conduct tactical diagnosis.
Sales decline on a national level can be identified through a hierarchical approach that helps determine if there’s one particular region with an isolated issue or if the issue applies to all markets as a whole.
The true power of using attributes and hierarchy together is realized when the two work in conjunction to determine how to respond to noise at different levels of granulation and to provide insight into what may be adjusted as a result of the identified trends. For instance, by identifying the decline in margin as a result of a particular brand, you could then look at the relevant attributes such as the depth of promotion or retailer type; however, the investigation is applied to determine why there has been a margin decline and what could be done to address it.
There are two sides to the equation; attributes support optimization for decision making while hierarchy provides a framework for prioritization. You need both types of information when making decisions. Marketing analytics is not just a means to report performance results; rather, they provide a logical method for analyzing the way the company does business.
When attributes are well defined and hierarchy is clean, analytics move from reporting on what happened in the past to a guide on making strategic decisions.
