Complex issues demand complex descriptions. Models are the representation of customer decision making and behavior --- these are complex issues.

At Tedesco Analytics models provide the backbone for understanding customers. We employ models and simulations for prediction, segmentation, strategy evaluations, and visualizations of emergent behavior.

The operations of the marketplace cannot be simplified or reduced. Traditional analysis based on linear equations or curve fitting only begin to represent the relationships between market factors and customer behavior. (see our paper on complexity models for the marketing mix)

A marketplace in motion is one of the most prominent, of the many, advantages we hope to realize by thinking in the complexity perspective. Our modeling efforts in consumer behavior move to reflect the living nature of decision making. Just as the specifics of economic modeling are changing from a mechanistic point of view to a new point of view, which actually described as organic, so too should the models of marketing reflect this change.

  Models are a quantitative and dispassionate view of marketplace behavior. The use of a model to simulate alternative strategies is the cutting edge approach to growing brands.

 

 

 

Complexity science explains systems that:

  • Display significant changes from small movement of initial conditions
  • Experience feedback from system components
  • Are non linear and with interactive elements
  • React with their environment Tends to self organize

The consumer and the business customer as a market exhibit all of these fundamentals. To fully understand this market behavior we must use a structure that embraces these dynamics. Complexity science is the answer and, models are the currency of this science.

 

Characteristics of Neural Networks

  • Discovers relationships present in data
  • Can apply previously learned information to new situations
  • Finds subtle patterns of casual behavior
  • Ignores irrelevant data
  • Effective for large data sets
  • Able to combine inconsistent data from a variety of sources and time periods

 

Neural networks are the primary modeling technique for use in applying complexity science to marketing. (see our 1992 paper ÒNeural MarketingÓ)

 

 

 

Project Applications

 
  • Forecasting and prediction
  • Market segmentation
  • Marketing mix optimization
  • Data mining and knowledge generation
  • Media threshold estimation

  • Market evaluation grids

 

Consumer behavior simulation

 

  Neural networks are comprised of processing nodes which can be layered to organize the representations of a marketplace. Layers are used to input data, compute interactions, and create predictions. A number of our papers provide detail on the specifics of neural network operation, application, and structure.

  The arrangement of neural network processing units, or neurons, is driven by the purops of the given model. In this example, a model is constructed to understand the demographic and socioeconomic profiles of magazine readers.

  Using a model, simulations are conducted which indicate the impact of various drivers within the context of interaction in the marketplace.

 

Self Organizing Maps
Guide to the consumer landscape

 
  • Permits visualization of consumer commonalties
  • Develops groups of consumers
  • Preserves the interactions among variables

 

Segmentation analysis is accomplished by a class of neural
networks know as self organizing maps.

 

 

Artificial Life

 

   Artificial life simulations (download our paper on the subject) provide insight on the interacting affect of marketing decisions.  

 

Visualizations

New modeling methods emphisize the added value of visual analysis. Temporal data reveals patterns in behavior that bring a deeper understanding of consumer decisions.