Modeling marketing data to improve performance

Gartner predicts that, by 2017, 33 percent of Fortune 100 organizations will experience an information crisis due to their inability to effectively value, govern, and trust their enterprise information[1].

The current explosion of data and information, particularly due to the ever growing impact of information and communication technologies, and the related dramatic changes in consumers’ behavior, is not comparable to what “normal” companies used to experience just two decades ago, and the gap between the amount of available data and the companies’ ability to manage them is growing exponentially.

This is not just a problem for large companies, but affects smaller organizations as well, and is even aggravated by their lack of familiarity with advanced information management systems: making appropriate and timely business decisions, based on the huge mass of available data and information, is becoming more and more difficult for everybody.

On the other hand, decisions are made for the future, and precise data about the future do not exist by definition. The management’s task is therefore that of identifying and distinguishing:

  • the variables that are totally outside their span of control
  • those that are somewhat controllable
  • those that are only indirectly controllable
  • overall, cause-effect relationships among variables that could better help at predicting, although with an inevitable degree of uncertainty, the likely future outcomes of specific combinations of exogenous and endogenous factors and decisions.

Unfortunately, on top of the information explosion problem, the industry and market phenomena that managers need to analyze and interpret have always been characterized by a series of interrelated features that make this task particularly difficult and complex.

One the most important of these features is the non-linearity of the relationship between investments (any type of investment: e.g. advertising expense, size and quality of the sales force, resources allocated to R&D, etc.) and results specifically attributable to these investments (e.g. awareness of the company’s brand, number of orders made by clients, quality improvements, etc.):

  • below a given “minimum level” of investment (normally called “threshold” or “critical mass”), the expected results do not materialize
  • beyond this level, results grow rapidly, up to a maximum level of investment, i.e. a sort of “ceiling”
  • beyond this ceiling, results could possibly improve further, but at diminishing rate, and therefore additional investments would not be convenient.

Obviously, the “critical mass” necessary to compete vary greatly, depending on the specific business, the type of investments and expected results, and the market and competitive context: within a given context, management needs to figure out, at least, which are the “reasonable” critical masses of specific resources necessary to successfully compete, and then allocate investments among these resources according to the estimated relative effectiveness and efficiency of each resource in relation to the components of value perceived by the market.

We will come back to this important issue after having discussed the other typical features that characterize industry and market phenomena.

[1] Quoted from Information Builders (2016), Driving Better Business Performance with a Practical Data Strategy: