Advanced Analytics Model Review & Validation

If Most Operational Spreadsheets Contain Errors¹, What Percentage of Your Advanced Analytics Models Contain Errors?

Since the consistent prevalence of spreadsheet errors became accepted, many organizations have adopted or procured formal testing and auditing for their spreadsheets. However, in advanced analytics, there has been no similar establishment of a quality assurance process—despite the commonality of error rates across all types of large, operational software applications. Typical software QA and testing methodologies are challenged to accommodate the special needs of advanced analytics applications.

We believe leaders of advanced analytics teams should be much more frequently asking, “How do we know our models are correct, and who has validated them?” Based on our work in applied optimization, Princeton Consultants will validate and improve your optimization and predictive models through our expert third-party review, and help you answer vital questions such as:

What is a correct model?
What data is being integrated and how?
How are solutions published and used in the business?
What validation was done to see that the model met the needs of the business?
How sensitive are the answers to the inputs?
How often is the quality of models assessed, considering the changing nature of the data and the business?
Did the implemented model reflect the intentions of the practitioner?

Our model review and validation will benefit your advanced analytics practice—and help you contribute even further to lines of business and their competitive advantage.

Uncover opportunities for improvement
Benchmark your models and group against best practices
Give business leaders more confidence in your solutions
Help your specialists improve their development skills

We answer more detailed questions specific to optimization and predictive analytics. Read more here.

Engagement Summary

Princeton Consultants blends technical and management consulting expertise to conduct a rigorous review and validation. Our process consists of the following steps:

Interview stakeholders from business and analytics development to understand business problem and context
Review of existing models and procedures
Review of data sources·
Implementation of models in alternative technologies to compare results (languages, solver, analytics engines·
Experiments with models to uncover issues

Who We Are

Princeton Consultants is an information technology and management consulting firm with offices in Princeton, New Jersey and New York City. Founded in 1980, we have completed over 1,500 projects for our clients--many of the largest, most successful and innovative companies in the world. We are a member of the INFORMS Roundtable, a group representing leaders of applied advanced analytics.

We help businesses in Transportation and Logistics, Financial Services, Media, and diversified industries to dramatically improve performance by integrating advanced analytics and new technologies into their core processes. Through our proven methodology, our clients have improved organizational decision making and experienced extraordinary gains in productivity, efficiency and customer service.

1. "88% of the 113 spreadsheets audit contained errors... A very rigorous testing stage after the development stage is needed to reduce error rates."
"What We Know About Spreadsheet Errors"
Raymond R. Panko