Approaches to, and even definitions of, an enterprise data warehouse (EDW) vary based on an organization’s architecture and application landscape. Accordingly, as you formulate an EDW strategy, you need a methodology that is well-defined and tailored to your organization’s needs. First, you have to understand the motivations behind creating an EDW. Next, you should assess your organization’s skills and financial and technical capabilities. From the results of these two initial efforts, you can then define and prioritize goals for your EDW program.
If you have already started your EDW initiative, use the guidelines in this article to validate and reinforce your program. If you haven’t begun your initiative, start now!
Understand the Motivating Factors for Establishing an EDW Strategy
Companies choose to introduce an EDW for various reasons, including to:
- Enable a single source of truth across the enterprise
- Implement data harmonization for incompatible operational data models
- Centrally manage the quality and consistency of information
- Provide historical “corporate memory” for analytical and compliance use cases
- Centrally manage the definitions for enterprise metrics, such as key performance indicators (KPIs)
- Create a consistent and comprehensive source of information for other applications
- Address TCO and time-to-value issues resulting from multiple data warehouses, databases, and data marts
While this is not an exhaustive list, it does capture many of the major themes usually included in a discussion of whether to establish an EDW. These reasons are also not mutually exclusive; they can be used in combination when constructing the initial business case for an EDW program or project.
Depending on your organization’s level of business intelligence and EDW maturity, it may be appropriate to estimate the benefits from all seven of these areas and then focus on the ones that are most urgent and that provide the greatest quantifiable benefit.
Assess Your Ability to Deliver
Once you have examined and started to quantify the benefits of creating an EDW strategy, I recommend getting a realistic overview of the issues that could impact your ability to deliver on that strategy. These issues typically fall into three main categories: financial, technological, and training.
Key Financial Issues
Financial considerations include central vs. decentralized IT budgets, ROI requirements, and overall budget regardless of ROI. Again, this list is not exhaustive, but rather illustrates that financial considerations are often grouped into two categories: who is going to pay for the project, and how to measure its benefits.
If you have a centralized IT organization, it’s important to understand how IT’s performance is measured. If IT is decentralized, you need to know how the departments’ budgets are allocated and to what extent technology decisions are governed by the departments and lines of business. It is also critical to understand if and how IT gets credit for ROI and benefits that accrue to the rest of the organization. Finally, you should be aware of the estimated total budget for the EDW strategy and how ROI will be measured — regardless of who controls the purse strings.
Key Technological Issues
The next category involves the existing technology environment. Even enterprises without an EDW strategy tend to have related tools already in place, which typically address one or more of these EDW layers:
- Data provisioning. Relevant technology includes extraction, transformation, and loading (ETL) tools, data quality products, and metadata management tools.
- Data persistence. Related technology includes RDBMS, OLAP, and prepackaged data marts.
- Information consumption. Reporting tools, dashboard tools, and spreadsheets are applicable here.
Understanding the capabilities and scope of the tools you currently own helps ensure the greatest possible reuse while identifying gaps and deficiencies that you will need to fill elsewhere.
Remember that your EDW strategy has a much greater chance of succeeding if you can build on what already exists rather than start from scratch.
Also be aware of any existing technology standards or approved vendor lists that could influence or limit your options. While it is often possible to secure exceptions, it’s a good idea to understand what choices have already been made and why. It is also helpful to know what standard and approved interfaces are available for data acquisition in order to determine the appropriate technology for loading data into the EDW. Any known interfaces for exporting data from the EDW to provision other applications should also be included in the analysis so you can fully comprehend the requirements and limitations.
Key Organizational Issues
The final area of concern is organizational and training issues. This often-overlooked group has the greatest potential to derail an EDW program. These issues encompass the skills and abilities of existing resources, internal politics and preferences, and organizational limitations.
It makes sense to be aware of the skills and abilities of the people who will build and maintain the EDW, but it’s also important to understand those of your end users. This will help determine typical consumption patterns and allow you to choose the appropriate tools and applications to facilitate these patterns. Be sure that your strategy not only leverages the existing skills and strengths of your organization, but also addresses the weaknesses.
Internal politics and preferences are also often underestimated. Whenever possible, you should make these issues explicit and address them directly to help eliminate any hidden agendas that can sabotage your EDW program.
I recommend addressing organizational and political problems on an ongoing basis in order to increase both the value and sustainability of your EDW strategy.
Set Appropriate Goals
Once you understand the benefits of creating an EDW strategy and acknowledge your limitations, you can create reasonable and attainable short-, intermediate-, and long-term goals. The following tips can help you set the appropriate goals for your EDW strategy.
Developing Short-Term Goals
- Choose attainable targets for domain and application scope based on stakeholder commitment and technical requirements.
- Don’t be too ambitious. Short-term goals should focus on gaining support and awareness.
- Look for tangible, clearly measurable benefits.
Set at least one short-term goal that will be highly visible across the enterprise.
Developing Intermediate-Term Goals
- Pick an application or information domain that can be completely integrated into the EDW.
- Look for opportunities to rationalize and eliminate information silos.
- Plan to mitigate or eliminate challenges that may arise from your short-term goals.
Set specific goals for evangelizing and communicating EDW capabilities and successes.
Developing Long-Term Goals
- Plan to roll out the EDW to all areas of the business.
- Try to enhance EDW data with external sources.
- Create agile EDW extensions for greater self-service and data exploration.
- Implement appropriate service level agreements and governance processes.
With these guidelines in mind, you can limit inherent risks and greatly increase the likelihood of success for your EDW strategy. Adopting and integrating an EDW requires far more than a purely technology-focused approach. Also be sure to review and revise the strategy regularly to ensure that it addresses changes in your business environment and leverages the latest technical innovations.
For more information, please visit www.sdn.sap.com/irj/sdn/nw-bi.
Jake Klein (firstname.lastname@example.org) is Vice President of Data and Analytic Engines solution management within the SAP Intelligence Platform and SAP NetWeaver Organization. Jake’s responsibilities include product strategy, definition, and solution rollout activities for SAP’s business warehouse, business warehouse accelerator, and enterprise portal portfolios. Jake is located in Ra’anana, Israel.