I am often asked usually by programmers - What is Data Warehousing & how do I learn it? I explain to them we use all the same tools that you do but differently. That’s when I coined the term Data Sense. It describes the essence of Data Warehousing and separates Data Warehousing from rest of Programming. Every aspect of IT from Hardware / Software infrastructure to Design, Development and QA is done with massive data flows and need for data precession accuracy and meaning.

Sunday, February 18, 2007

Business Situations & Approaches : Examples

Customer: Billion $ Telephone Company Newark, NJ

Project Name: Enterprise Data Architecture and Data Warehouse

Business Situation:
Client wanted to replace ad hoc stove pipe data marts and application databases in the company with a more structured solution.
The current solution is manual and resulted in exploding resource requirements and dipping user satisfaction.
A massive one-time rewrite is ruled out as impractical.

Approach:

Design the ideal technical and data architecture and standards to be used though out the company.
Select best of breed software packages tools and applications as company standard.
Rewrite and move applications and data marts one by one to the new architecture making sure they integrate with existing solution.
Selection of a mid-tier storage platform to provide a consolidated storage platform for all applications and data marts.

Solution:
· Linux grid with blade based architecture to help reduce footprint of server racks and provide flexibility, expandability and redundancy. Software tools selected that work in grid environment.
· EMC SAN for storage of more than 10 TB of data.
· After full implementation this project will cut the staffing requirements by over 50%.
· The turn around time for new projects down by two thirds.
· Users are more satisfied with the accuracy and the timeliness of the data.
· Company cut down by over 80% the money not billed and collected because of the faulty systems.


Customer: Harmonic Corporation, San Francisco, CA


Project Name: Common Advertising Platform
Business Situation:
A new advertising company was focusing on building a common advertising platform for multiple media. From TV to newspapers and Internet.
Need an end-to-end solution to support the business with a new idea and requirements.
Need to integrate with the current software facilitating the advertising in various media and various vendors.
Approach:
Understand each of the existing business processes in various media and develop a business model for the common platform to support all corresponding tasks.
Understand interfaces to the existing solutions in various media that the common platform has to interface to.
Develop an object model and a relational data model to support the common business model designed above.
Develop a BI strategy to deal with multiple clients and multiple media.
Develop a technical and data infrastructure for BI and interface with common advertising platform’s infrastructure.

Solution:
This is a new application that initiated the company’s endeavors in this area.
This application was latter sold as independent solution.
Solution developed on Sun SOLARIS with EMC as storage.
Java is used to develop the applications on ORACLE. Industry best tools selected for BI.


Customer: PetSmart, Phoenix, AZ

Project Name: CRM
Business Situation:
Company had flat sales over last few years.
Increase sales by using CRM tools.
Introduce Loyalty Cards
Company had no consistent idea of its customers.

Approach:
Did business analysis to discover nearly 50 different stores of customer information. Initiated an effort to standardize the customer information collection. Designed and implemented a customer database.
Short listed CRM vendor solutions for their fit against business requirements and IT Standard and Architectures.
Selected vendor based on Proof of Concept and total score within Vendor Selection matrix tool.
Design and implement a Loyalty Card program using the CRM tools implemented.

Solution:
The CRM solution was found to increase the sales in target market by over 10%.
Company’s standardized customer database was used for several other initiatives.
Oracle on HP UNIX was used with EMC as Storage and ORACLE as database. Code was developed using PL/SQL.



Customer: WebMD, New York, NY

Project Name: BI Migration
Business Situation:
The client discovered that with the dot COM crash costs need to be cut by over 80 % to stay alive. Every system was migrated from existing platform to a simpler cheaper architecture with an entirely new group of engineers.
BI fell through the cracks. There was no solution 2 months before scheduled live date (latter postponed for unrelated reasons)
With little budget, and no existing knowledge base to draw from a crash effort is in place to rewrite the old BI solution to new platform

Approach:
Technical infrastructure to be similar to rest of solution to take advantage of existing resources.
Quick business analysis by reverse engineering requirements from the existing solution.
Simple Data Architecture following the established principles in industry
Rapid application development using two dedicated resources and a number of short-term contractors as needed.
Move requirements where necessary to next phase of project.

Solution:
Solution was ready for launch with rest of the applications.
The team size was reduced by over 75 %.
The company saved over $1 MM per year in software and infrastructure costs. The performance increased by a factor of 4.
The solution was developed on SQL server on windows 200. Hardware storage solution is home grown using RAID disk array.