Reliable predictions yield confident decisions.

White Paper:

Modeling Best Practices

Proven methods to improve marketing demand models

White Paper:

Marketing Mix Modeling 101

How to measure which individual marketing activities impact consumption

Modeling Technology and Services

M-Factor develops sophisticated forward-looking demand models that allow you to continuously analyze, forecast, and optimize marketing investments and trade spend.  Our “mass-custom modeling” approach enables the delivery of high quality models in a fraction of the time of traditional modeling techniques without relying on “industry standard” cookie cutter approaches that fail to capture the nuances of your particular business.

Next Generation Modeling Technology
M-Factor’s model developers use our proprietary modeling software.  Based on state-of-the-art statistical estimation algorithms, it offers three key advantages over standard statistical packages:

Designed for demand model estimation; operates within the customer analytical environment, provides full transparency of the modeling process.

Designed specifically for demand model estimation

  • Allows early detection of abnormalities in the data and results and quick modification and testing of assumptions — greatly improves modeler productivity

Operates within the customer analytical environment

  • Provides relevant context for evaluating the models in business terms rather than only based on their statistical merits — greatly increases model utility

Provides full transparency of the modeling process

  • Allows M-Factor modelers to work interactively with customers to incorporate specific business feedback — greatly increases model quality

Better Predictions Come From Updated Models
Most demand models perform poorly out of sample because they are designed for retrospective analysis and are not updated to reflect current market conditions.  This is why the “other” or “unexplained” bars (euphemisms for model error) in “Due-To” reports grow over time.  M-Factor has been awarded a patent on tools and methods it developed to update models, and consistently generates models that are more predictive.  In a recent benchmark example, M-Factor models empirically reduced out of sample error by 68%, with .99999 confidence.

© 2004–2010 M-Factor Inc. All rights reserved.