• // Business Intelligence
  • // Data Warehousing
  • // Business Analytics

5 Steps to Solve the EDW Puzzle: Consolidating Line of Business Marts

Part of the evolution of a company’s business intelligence (BI) maturity is the ability to implement a true enterprise data warehouse. But, as we’ve often seen with our business intelligence consulting clients, that is not what the companies typically start with, for cost and time reasons. More likely, in an effort to realize faster BI ROI, companies will undertake smaller, focused BI efforts, and before companies realize it, they end up creating multiple disjointed BI stacks across the organization serving specific lines of business.

The Cost of Line of Business Marts

While these silos may serve the needs of a line of business (LOB), they come at a cost when it comes to cross-LOB or enterprise analytic and reporting needs. A lot of manual crunching, number consolidations and reconciliations across systems is done and as a result more home grown, DIY analytics and reporting solutions take shape. For enterprise analytic and reporting needs, it means no visibility or consistency of data across the organization. Unfortunately, more often than not, there is no easy means to bubble up the information for enterprise BI capabilities from these BI silos.

Five  Steps for EDW Consolidation

There is no one-size-fits-all solution for creating an EDW from a state of LOB marts, nor is there an easy solution to bringing them together. Here’s an outline of five steps you should consider as you look to consolidate your LOB marts. Of course the process is significantly more complex, but these overarching questions will help guide the consolidation process.

1. Determine the Business Need

Needless to say there has to be a business need to bring them together and to create synergies out of the existing marts. So the foundational and the most critical work revolves around getting to know from business what their information needs are, what they are able to achieve currently with their LOB data marts and what they would like to be able to achieve to run their business better. This is where you would ask them about their home grown analytics or number crunching and consolidations they do outside the formalized marts. This helps to put the gaps in their information needs into perspective and provide insight into whether bringing information together would be an ideal state for information delivery and consumption or not.

2. Uncover the BI Opportunities

The enterprise information needs that are gathered in the previous step can be categorized into business intelligence opportunities (BIOs). Examples of BIO would be workforce management analysis, customer analysis, risk analysis, etc. These BIOs can be prioritized by the business for implementation in phases. Implementing the BIOs that would generate the biggest bang for the buck and also completed in a shorter timeframe would be ideal to showcase the success of the EDW efforts.

3. Assess Your Technical Infrastructure

An effort to gather the current state of the technical BI environment, business served, the environment, tools used and the infrastructure etc. should be conducted. The performance and efficiencies of the existing technical BI architectures and environments should be gauged for maturity and conformance with best practices.

4. Build a Technical Implementation Roadmap

A lot of considerations need to go into building this roadmap, some being determining the platform for target EDW, evaluating the tools needed and types of tools needed, evaluating a data architecture that would define the EDW and the follow-through implications from that architecture; deciding whether to virtualize the EDW etc. Sometimes an objective evaluation of these considerations could be done by utilizing a multi-attribute model.

5. Last but not Least: Manage Data Governance

The success of your EDW is only as good as the quality of data in it, so data governance is crucial. It helps build confidence not just in the numbers, but in the validity of BI projects, and your overall BI/DW program. In short, data governance serves as an over-arching monitor at each stage of the EDW process. I talk about it in more detail in my previous post: Get Proactive about Data Quality.

Remember, Getting to an EDW is a Journey

Finally, you cannot wake up one fine morning and have an EDW. It’s a journey, and with each phase there will be more and more business value realized from the EDW. BI is no longer a choice for a company, it has become a norm and the phased, patient and diligent efforts to build an EDW will definitely pay off in the short and long run.

by Balakrishna Dixit, Principal DW Consultant

© DecisionPath Consulting, 2011

Created by Matrix Group International, Inc. ®