Information Technology – Global Services

Case Study

Market Situation

Prior to Modern Analytics involvement, the client built 150 individual predictive models offshore using 35 statisticians. Delivery timeframes were weeks and months and the costs in line with more traditional, hand-crafted methods.

Background
  • Our client serves over 5 million business and 20 million executives across 200 countries
  • The client delivers 50 major products and services and hundreds of applications and software platforms
Objective
  • Personally nurture each executive with custom communications and education
  • Improve executives’ ability to provide value added solutions to the business
Strategy

To improve the efficiency of their marketing, communication, and sales resources, the customer determined they needed to vastly increase their ability to “predict” the likelihood of the following:

  • A particular business or executive’s receptivity to a particular message though combined communication channels
  • An executive’s increased purchasing power or modifications to their budgets
  • Shifting receptivity at both the company and executive level, much like a consumer credit score

Solution

Using Model Factory, we fully automated the entire process of asking questions, creating predictive models, getting answers quickly and putting them into production. The first-year goal was to scale from 150 models to 350 with a second-year goal to produce and maintain 2,500 individual predictions and then to scale into the many thousands as users adopted the new methodologies and technologies.

Today we are generating, with fewer than 2.5 FTEs, more than 4,800 individual predictions. These predictions produce 11 billion scores monthly supported by 5 trillion pieces of information. With machine learning and automated governance, we deployed tournament style predictive analytics. We exhaustively search every known data variable, and deploy all major modeling techniques in a fingerprint-free environment.

The customer is looking to increase the scale several-fold.

Results
  • Targeted high-value customers with personalized communication and custom recommendations
  • Enhanced customer experience and satisfaction through prescriptive data analysis
  • Streamlined the analytical process and eliminated time-consuming manual steps with automation
  • Reduced staffing requirements from 35 to fewer than 2.5 FTEs
  • Generated 4,800 individual predictions and 11 billion scores monthly