StatWizards develops and uses advanced mathematical tools to solve critical business problems.  The techniques we employ include:

  • Discrete-choice models
  • Latent-class segmentation
  • Machine learning
  • Predictive modeling
  • Fader-Hardie models
  • Econometric models

A list of models in itself is not interesting.  What is interesting is the list of business problems these models can address.


From a business perspective, the family of discrete-choice models (DCM) represents the decade's most significant breakthrough in quantitative business tools. By analyzing the way in which customers make choices, either in a controlled or real-world setting, DCM can infer how customers trade off product features and pricing in a competitive environment. Why is this important? Because the right combination of features and pricing can improve product positioning and add millions of dollars to a firm's bottom line.  DCM is often used in new-product introduction, in which companies want to fine-tune product positioning.


Not all people want the same things from products.  The notion behind segmentation is that by understanding differences in consumers’ needs, businesses can design product variations that more closely meet those needs.  Latent-class segmentation is arguably the most effective way of classifying people into market segments.


As data collection methods improve and storage costs decline, companies are accumulating extraordinary amounts of data about their customers.  Big data techniques such as machine learning help companies glean insights from this data and transform them into recommendations for action.


A comparative newcomer to the quantitative toolbox, Fader-Hardie models comprise a brilliant set of tools that can address important business issues.  Their general objective is forecasting, but Peter Fader of the Wharton School and Bruce Hardie of the London School of Business have applied their ideas to a range of business applications, including:

  • Identifying underlying customer segments
  • Calculating customer lifetime value (CLV)
  • Forecasting contract renewal rates
  • Projecting customer attrition
  • Forecasting trial and repeat sales of new products
  • Understanding the drivers behind the frequency of visiting a web site
  • Determining the optimal size for a mailing based on a test mailing
  • Generating effectiveness measures for a media exposure

Their breakthrough came in developing a way to simultaneously estimate individual processes (like buying processes) and their distribution through a population. The resulting tool forecasts at both the aggregate and in most cases at the individual level, and does so with remarkable accuracy.


Many products are subject to seasonality and/or business cycles.  Understanding this relationship can help firms match production to demand, calibrate inventories and time promotional activities.  Given a data set of historical sales, we build models that identify key causal economic variables and predict demand over time.


TURF stands for Total Unduplicated Reach and Frequency.  The method addresses a common business problem: with a limited advertising budget or shelf space, where should we advertise or which products should we display.  By using an advanced way of counting readership, viewers or product preferences, the technique determines the mix that reaches or appeals to the most customers.