Insurance Industry Business Intelligence and Data Mining Services

Business Intelligence Solutions can help insurance companies, third party insurance administrators, state insurance funds and state regulatory agencies in the following five primary areas: actuarial, CRM and marketing, underwriting, claims analysis and network functioning analysis through the usage of the state-of-the-art simulation/optimization, data mining services and GIS technology. Business Intelligence Solutions can help you answer any specific data driven business question.
Definition of data mining and detailed characteristics of a data mining project can be found in our white paper “Data Mining: the Means to a Competitive Advantage”.

In order to understand why traditional statistics is not always adequate tool for vast majority of data driven  insurance industry problems, see our presentation "Data Mining vs. Statistics". 

The essence of our approach is to understand and analyze our client’s business problem and corresponding data through the prism of diverse statistical/data mining models.  As a result we are always able to produce the best possible model /results and help our clients in the most effective and scientifically sound way. In particular, our business intelligence data mining consulting services can help you solidify your risk-based pricing, detect potentially fraudulent claims, analyze and model P&C policies and claims data to estimate insurance risk and discover previously unrecognized risk groups in order to help actuaries develop more effective, more accurate, and more competitive rating systems that better reflect the true risk of underwritten policies.

Workers Compensation Insurance Risk Management

Claimant fraud and provider fraud detection (independently or with cooperation with SIU), optimization of claims management, early assessment of claim severity, predictive modeling and forecasting of claim attributes (claim duration, medical and  total cost, disability duration,  actual claim settlement value, etc.), claim cost reduction management and saving quantification, claim scoring and  drivers identification of becoming severe claim, cost of claim, etc. Evaluation of effectiveness of treatment, using claim data, claimant attributes, and injury/disease/treatment information, return to work management optimization.

Health Insurance Risk Management

Analysis of hospital claims, Predictive analytics, modeling and forecasting of hospital cost and length of stay, identification of influential factors/drivers of cost and length of stay, analysis of admission/discharge data, analysis emergency room data, patient flow management.

Network Functioning and Network Structure

Analysis and modeling of employer decisions to purchase health insurance services, purchasing decision drivers, identification and quantification, likelihood of purchasing across different regions and industries, developing programs to increase purchasing probability, scoring of employers in terms of the likelihood of purchasing, switching to competitor services, profitability etc. Employer segmentation and profiling. Determining the most profitable segment of employers. Employer targeting strategy development. Identification of underserved employers (employers that did not purchase your services, but which look like potential purchasers – these employers are most likely to purchase your services in the nearest future if the corresponding marketing program will be designed and implemented).

Analysis and modeling of employee decisions to select a particular plan

Drivers identification of employee decision, profiling of employees that selected each plan, identification of importance of different attributes in employee decisions.

Plan attributes (premium cost, method of compensation, freedom/flexibility to choose a doctor, control of treatment decision, diagnostics coverage, etc.).
Network attributes ( network size, ratio of specialists to generalists, access factors, utilization / performance measures, etc.).
Employer attributes ( industry, geography, number of employees, years of usage of particular services, etc.).
Employee attributes (marital status, number of dependents, age, education, preconditions, years of usage of the plan, etc.).

To improve the predictive accuracy of models, employee’s neighborhood characteristics, such as demographics, socio-economic status of the neighborhood, community tapestry segment, market potential, lifestyle variables, etc. can be incorporated into the analysis.

Analysis and forecasting of network voluntary and network involuntary turnover rate. Impact of marketing events on service sales and network structure and functioning. Market positioning to understand your product from the customer’s point of view relative to the competition. Segmentation and profiling of a plan’s affiliates, and in particular, determining the most profitable segment of practices, and/or network affiliates by specialty /  geography  / employer / plan. Development of a program to increase profitability. Identifying areas with insufficient network coverage and large demand of your services. Identifying areas with sufficient network coverage and small demand of your services.

Customer Relationship Management

Gaining, sharing, expanding, and utilizing knowledge about customers. Customer profiling, customer segmentation and clustering with the usage of demographic, geographic, psychographic and behavioural variables and historical business information (see our presentation “Cluster analysis vs. Segmentation”). Data enrichment (the essence of data enrichment and different sources of external supplementary data for CRM problems are discussed in our presentation “Data Enrichment for CRM”). Improving marketing ROI by targeting specific customer segments with personalized campaigns.

Developing lifelong relationships with your most profitable customers and addressing other CRM challenges like up-sell and cross-sell (application of bivariate regression to up-sell and cross-sell is described in our white paper “Joint Regression Model for Sales Analysis”)

Identification of early adaptors. Customer profitability and customer loyalty. Customer acquisition and retention. Churn modeling/customer defection and customer attrition. Customer satisfaction, identification of dissatisfied customers, and development of programs to improve customer satisfaction. Net Promoter Score analysis.  Personalized marketing and pricing. Targeted marketing and multichannel management. Optimal (the most profitable) allocation of your marketing funds. What-if scenario analysis. Detecting underserved customers and unprofitable customers and developing program to increase sales. Customer scoring according to a likelihood of defection, likelihood of response, likelihood of a product purchase, etc. Developing lifelong relationships with your most profitable customers and addressing other CRM challenges. An example of customer scoring can be found in our case study “Data Mining Approach to Credit Risk Evaluation of Online Personal Loan Applicants” presented at Fifteenth (2009) ACM SIGKDD International Conference  on Knowledge Discovery and Data Mining in Paris, France.

Customer Knowledge Management

Gaining, sharing, expanding and utilizing the knowledge residing in customers (e.g. through survey development and analysis of combined primary and secondary data).

Design and Execution of Specific Campaigns

Targeted marketing to optimize marketing effectiveness, customer acquisition, cross-selling, up-selling, retention campaigns, etc. Customer win-back campaigns (recovering customer lost to competitors). Pilot study design (e.g., price test to estimate customer sensitivity to a product/service rate increase). Event monitoring (e.g., when a customer's account total reaches a certain dollar volume) and event prediction using predictive analytics.

Web Analytics

Website analysis, predictive analysis and predictive modeling of website user behavior (e.g. prediction of user propensity to convert, buy or churn), analysis of website statistics and key performance metrics, measuring content effectiveness, discovering user segments and navigation patterns, customer satisfaction with web experience, uncovering areas of opportunity for your organization’s websites to more effectively support sales and customer loyalty. Impactful e-mail marketing. IP address geocoding of website visitors to identify geographic clusters of current and potential customers; examine the efficiency of advertisement as a function of geography; determine locations of future advertising campaigns, etc.  In order to get insight into spatial dependency and spatial regression, see our presentation “A Brief Introduction to Spatial Regression”