• Re-engineer all business lines (loans, deposit products, home equity, investment products etc.) to better meet consumer preferences while lowering cost-to-serve
• For each business vertical:
  • Identify "trade-offs" attributes, drivers we that are NOT differentially important.  De-service those attributes in order to lower the overall cost-to-serve.
  • Identify new delivery configurations that will grab preference share for non-customers.
  • Identify blunting strategies; strategies that blunt competitive encroachment.
  • Identify optimal service configuration that will retain consumers.
  • Identify the most distinctly different segments in each vertical and the best ways to deliver service to each segment at the lowest cost.
  • Re-engineer the go-to-market strategy for each vertical.

• Meeting the actual "buyer values" of consumers grows targeted customer capture (acquisition) by 20%, deepens the dollar volume relationship an average of 10% per business line, and decreases attrition 30% bankwide.
• Identified "trade-offs" attributes, drivers we that are NOT differentially important.
• Tested hundreds of thousands of new configurations in the simulators and identified:
  • New delivery configurations that will grab preference share for non-customers.
  • Optimal service configuration that will retain consumers.
  • Competitive blunting strategies that seize maximal share from new "encroachers".
• Identified the most distinctly different consumer segments in each vertical and the best ways to deliver service to each segment at the lowest cost.
• Re-engineered the go-to-market strategy for each business line.
• Main Segments: Price,/Fee, Speed, Service/No Errors, Relationship.

• Adaptive Pairwise Trade-off Surveying: reproduces the real-world trade-offs and learns from prior answers to customize the survey to each consumer.
• Advanced Algorithms Find What Is Truly Important: Hundreds of combinations algorithms, sub-algorithms, iterations, weighting of survey sections uncovers the "utilities" that result in the simulator making the same choices as the consumer.
• SegmentSolveTM : our segmentation platform which uses 13 separate algorithms to converge upon the most highly separated consumer segments.
• ProductSolveTM : is our goal-solving market simulator.  By capturing values (what's important to the consumer) we no longer need to create endless surveys.  Instead we create new products  in the simulator (by what-iffing or goal solving)  to discover which products  consumers will buy and which will maximize customer retention.
• PositionSolveTM: Visualizes the positioning of each competitor and discover how to move into innovative but currently unpopulated areas of demand.

• Gets Down to the Bedrock - the Real "Key Drivers": adaptive pairwise trade-off whittles down the drivers that yield real-world outcomes (sales,  etc.).
• Industry-Leading Accuracy in Quantifying Those Key Drivers: Thousands of  combinations algorithms, sub-algorithms, iterations, etc , are tested to converge on the optimal "utilities." These utilities, in turn, enable the simulator to make  the same choices as the consumer.
• Produces Strategies That Maximize Bang for the Buck: via our goal-solving simulator.
• You Get the Fishing Pole Not the Fish: By capturing values (what's important to the consumer) we no longer need to create endless surveys.  Instead we create new products  in the simulator to discover which products  consumers will buy and keep.
• Generates the Most THOROUGH Recommendations: Because our goal-solving simulator tests hundreds of millions of strategies to discover the ones that meet YOUR criteria (e.g. the feature changes that maximize revenue at lowest cost to implement).

• 10% higher lift on average than "best practice" software or consulting .  (See why by inspecting the technologies described in the next tab.)
• Deployed cost: 25% of the cost of traditional consulting approaches without the need to rearchitect the datamart.

UHP-DM is an R&D project fifteen years in the making.  It has yielded two linked applications; VISTAR and BigDataSolve:

• VISTAR: is our automated sub-datamart reduction engine that identifies the best predictor combinations to maximize the desired outcome (objective function).
• BigDataSolveTM: takes those variables and runs millions of models within that solution space.  It does so by creating a solution space that automatically rotates algorithms, sub-algorithms, iterations, criteria and transforms against the problem,  This typically results in millions of model combinations that probabilistically yields higher lift.
• Distributed Computing: UHP-DM uses distributed computing by turning regular laptops in a network into a virtual supercomputer.
• DELTA: a specialized form of validation (Differential Enhancement of Lift Through Arbitration) yields higher lift by finding subgroups with higher predictive performance.

The Differentiation of our UHP-DM (Ultra-High Performance Data Mining) practice is industry-leading.  In two simple ways.  It is:

• Way Less Expensive: Consulting firms still rely on a standard formula 1) Scaring their clients to death with what is (mostly) the truth 2) Convincing them to 2) rearchitecting their datastore and b) buy expensive closed-architecture analytics.
• In Contrast to Standard Consulting, UHP-DM Gives You What You Need Far Less Expensively...But With BETTER Results :  There are only four keys to ultra-high performance data mining. Our UHP-DM approach addresses them all:
  • Getting the Right Predictors / Cleaning Them Thoroughly (See prior VISTAR tab.)
  • Finding the Algorithm & Transforms That Predict Best (See BigDataSolve.)
  • Doing it Fast w/ Inexpensive Distributed Computing: (See BigDataSolve / DP.)
  • Validating Exhaustively...in a Way That Increase Predictive Fit: (See DELTA.)