Thursday, February 18, 2016

Dimensional Modelling for Insurance Industry


Insurance industry is an important and growing sector for business intelligence market. Insurance companies generate several complicated transactions that must be analyzed in many different ways. Insurance companies like GEICO help customers manage their risks by charging a fixed premium amount and providing a sum of money in case of unexpected events such as a car accident or medical emergency.

Insurance Industry base their business models around assuming and diversifying risk. Most insurance companies make revenues in two ways:  charging premium amounts in exchange of insurance coverage and subsequently investing those premium into markets to earn interest.


Within Insurance Industry, I have chosen GEICO and will be applying concepts of dimensional modelling concepts to evaluate the performance metrics for evaluating an insurance company.  GEICO which stands for Government Employees Insurance Company is an automobile insurance company based out of Chevy Chase, Maryland. It is the second largest insurance company in United States of America. It is wholly owned subsidiary of Berkshire Hathaway. Today, there are more than 22 million automobiles insured by GEICO which are owned by more than 13 million policy holders. GEICO follows a direct-to-customer sales model.

From the perspective of GEICO’s CEO there are certain KPIS(Key Performance Indicators). Few of them are listed below

  • Average Insurance Policy Size - Average insurance policy size of policies closed within measurement period. This KPI is most used for: Operational Excellence
  • Loss Ratio % - The ratio of claims to premiums. It may be calculated in several different ways, using paid premiums or earned premiums, and using paid claims with or without changes in claim reserves and with or without changes in active life reserves.
  • Claims Solvency %  - Insurance companies’ ability to pay the claims of policyholders.
  • % of Overdue Claims - Percentage of overdue claims.
  • Percentage of Sales Growth - Measures the amount of policy renewals and new policy sales over a set period of time
  • Net Income Ratio - Measures how effective organization is at generating profit on each of dollar of earned premium.


Based on the above KPIs, I have picked Claims Solvency % as the one on which we will proceed with our case study. Claims Solvency % refers to an insurance company’s capacity to compensate its customers in case of unexpected calamities i.e. in GEICO case can be car collision.

Performance Metrics that are important to evaluate GEICO’s performance can be

  • Total Revenues – Total amount collected from customers in exchange of insuring their     automobiles
  • Total no of Customers – Total Number of Customers insured by GEICO
  • Total Settlements – Sum of amount paid to customers in lieu of insurance bought by customers in case of car collisions
  • Total Insured Amount – Sum of amount which GEICO is liable to pay in case all customers are need to be paid


Dimensional Model can play a crucial role in helping GEICO’s higher management getting answers to specific questions. Such as how many customers were insured on a monthly, Yearly or daily basis. The level of granularity to be queried can be handled by using Dimensional Models. Similarly, Dimensional Models can help them query and evaluate the performance of specific insurance products or a group of similar insurance products clubbed together.

Given the surplus amount of data collected during the operations of a large insurance company like GEICO, Dimension Model can help in ensuring simplicity and optimized query platform for business users like the CEO. Dimension Model denormalizes the highly complex transactional database tables and clubs them into dimensions, hence ensuring simpler and speedy query of data.
Dimensional Modelling maximizes flexibility and scalability depending on changing user activities. Given the notion, GEICO’s CEO can track real time changes occurring to the customer data and base their evaluation on that.

Transaction-level star join schema  would provide an extremely powerful way for GEICO to analyze insurance claims. The number of claimants, the timing of claims, the timing of payments made, etc can be easily derived from this view of the data.

Below is a sample dimensional model which GEICO can adopt in order to evaluate their performance.

Thursday, February 4, 2016

Business Intelligence & Analysis Products Scan & Evaluation

What is Business Intelligence?

Business intelligence (BI) is described as "the set of techniques and tools for the transformation of raw data into meaningful and useful information for business analysis purposes".  BI, is an umbrella term that refers to a variety of software applications used to analyze an organization’s raw data.
Simply put, BI allows easy interpretation of large volume of data and take intelligent business decision based on the interpretation.

There are several Business Intelligence and Analytics tools and platform, available in the market, which makes taking an intelligent decision much easier.
However, one needs to comprehensively evaluate a BI tool, taking various criteria into account, before choosing it for an organization and invest in it. Criterias such as data visualization capability, scalability and so forth must be considered before choosing the best fit for the organization.
This blog discusses the top five most popular Business Intelligence & Analysis products used today and evaluate these products based on certain criteria and rate them based on weight of each criteria.

Criteria used for comparing BI tools
  • Data visualization
    One of the most important feature in BI tool is data visualization. This criterion is the capability of BI tool to represent data, to their audiences, in the most interactive way possible. The tool should use vibrant visual objects, images, charts and interactive dashboard having informative visualization. The visual representation of data should be such that, for even a non-technical users, it is easier to understand
  • Scalability
    This criterion is the ability of the BI tool to handle growing amount of data, growing amount of users and its potential to be enlarged. BI tool should be scalable enough to add additional functionalities based on the business requirement. Business settings are susceptible to change, a good BI tool should perform well when even when business settings are changed rapidly
  • Cost effectiveness 
    Cost is another important factor which needs to be considered while choosing a BI tool. This criterion is based upon return on investment or value for money, in other words, number of features provided by the BI tool for the price that is charged.
    The cost of a BI tool includes the following factors:
    • License costs
    • Cost of BI developer
    • Training costs
    • Support/Maintenance fees
  • Data Integration
    This criterion is the capability of the BI tool to handle multiple data sets, originating from multiple data sources. These sources include XML data, MS Excel data, MySQL or Oracle Database, Siebel, SAP database, flat files, etc. The variety of sources from where data is being collected is ever increasing, it is vital for a BI tool to be regularly updated, to support multiple data sources and format.
  • Customer Experience/ Ease of Use
    This is a measure of how easy it is for the users to get used to the product and produce some meaningful output. The users should require minimal training to use the BI tool.This criteria also includes help and support documentation provided by the vendor



Business Intelligence Tools


1. Tableau



Strengths:
  • Supports data integration from various sources 
  • Wide range of dashboard capabilities
  • Drag and drop easy to use interface
  • Row level data security
  • Cost is comparatively less than other tools
  • High quality customer support to quickly resolve any issues
Weakness:
  • Lacks a robust security system
  • Lacks support of APIs
  • Less support for custom modification and 3rd party plugins
  • Limited functionality for data mining

Product Analysis:

Data visualization
All features are dynamic, interactive and highly customizable. One of the best tool available

Scalability
Issues with large volume of data

Cost effectiveness
Tableau is free for students and commercial version cost around $2000. Training is provided free of cost

Data Integration
Can handle variety of data sets easily



2. MicroStrategy


Strengths:
  • Ability to generate enterprise reporting, dashboards and notifications
  • Ad-Hoc reporting is supported
  • Mobile application to allow user to create reports on mobile phones
  • Enterprise grade cyber security to provide object level security
  • Multiple data-sources are supported 

Weakness:

  • Visualization effects are not that good as compared to other tools
  • Not easy for business and non-technical users
  • Product cost is higher compared to other tools offering the same feature set
  • Structured data warehouse is needed for integration

Product Analysis:

Data visualization
Reports lack visual appeal, also takes longer time to generate

Scalability
Highly scalable, can handle data up-to 17 TB easily

Cost effectiveness
The cost of the tool is reasonable and is around $3000

Data Integration
Can integrate with excel, hadoop and other big data sets. Excellent data integration



3. QLIKView



Strength:

  • Quickest implementation time and exceptional performance in vital BI projects
  • Ability to easily handle complex ETL
  • Excellent tool for data discovery, can dive deep into data for analysis
  • Excellent visualization features and easy to use user interface


Weakness:

  • Not very scalable
  • Few tasks require steep learning curve, not suited to business users
  • Server deployment could be expensive
  • Not good for real time reporting

Product Analysis:

Data visualization
One of the best BI tools for data visualization. Closely follows Tableau

Scalability
QLIK is highly scalable, supports large data-sets and thousands of users

Cost effectiveness
QLIK costs around $4000 and $1000 per additional user

Data Integration
QLIK can work with multiple data-sets with ease, better than Tableau in this criteria



4. Oracle BI


Strength:

  • Supports Big Data
  • Good detailed user documentation
  • User friendly interface
  • Excellent customer support and training
  • Ability to analyze large data-sets in short amount of time 

Weakness:

  • Not easily customizable, requires significant amount of time for custom requirement
  • Lack support for multiple data-sources

Product Analysis:

Data visualization
Oracle is good in handling large sets of data, it was primarily developed to process large scale system of record requirements, it is still not good in data visualization

Scalability
Oracle has a scalable architecture, it is a enterprise level tool and it handles large amount of data.

Cost effectiveness
A major drawback of Oracle BI is its high cost

Data Integration
Oracle BI is primarily uses SQL databases and is limited in processing multiple databases


5. SAS BI


Strength:

  • Easy and fast reporting
  • Supports advanced techniques like prediction modelling and data mining 
  • Data integration is excellent
  • Huge market share
  • Support of other programming languages like R


Weakness:

  • Cost is higher compared to other BI tools
  • User interface is not so user friendly
  • No hardware architecture support

Product Analysis:

Data visualization
SAS has good amount of features for data visualization

Scalability
SAS is highly scalable, it can operate on multiple operating platforms like Linux/Unix, Windows, etc.

Cost effectiveness
Cost of using SAS is too high, servers cost around $8000 and $1700 is charged for any additional user

Data Integration
SAS can handle mupltiple data sources like MySQL, Hadoop, flat siles, etc.



Weight score analysis of BI tools


CriteriaWeightTableauSASQLIKMicroStrategyOracle
Data visualization30.00%1081066
Scalability15.00%8108810
Cost effectiveness20.00%88777
Data Integration20.00%88886
Customer Experience15.00%1081098
Points100.00%8.98.38.77.357.1
Rank13245


After carefully analyzing each BI tools and comparing them on the basis of Data visualization, Scalability, Cost effectiveness, Data Integration and Customer Experience. We see that Tableau is clearly a better tool than the other four BI tools. However, as we can see all the five tools are close enough to compete against each other and each of them has their own unique set of features, the choice of selecting a BI tools will eventually depend on specific set of features that the user is looking for


References:
https://en.wikipedia.org/wiki/Business_intelligence
http://www.qlik.com/
http://www.microstrategy.com/us/
http://www.tableau.com
http://www.sas.com/en_us/home.html
http://www.oracle.com/us/solutions/business-analytics/business-intelligence/foundation-suite/overview/index.html
http://www.cio.com/article/2439504/business-intelligence/business-intelligence-definition-and-solutions.html
http://www.gartner.com/technology/reprints.do?id=1-2ADAAYM&ct=150223&st=sb