Do not launch the data warehouse unless and until your company is ready for it.
Find the best executive sponsor. Ensure continued, long-term, and committed support.
Emphasize the business aspects, not the technological ones, of the project. Choose a business-oriented project manager.
Take an enterprise-wide view for the requirements.
Have a pragmatic, staged implementation plan.
Communicate realistic expectations to the users; deliver on the promises.
Do not overreach to cost-justify and predict ROI.
Institute appropriate and effective communication methods.
Throughout the project life cycle, keep the project as a joint effort between IT and users.
Adopt proven technologies; avoid bleeding-edge technologies.
Recognize the paramount importance of data quality.
Do not ignore the potential of data from external sources.
Do not underestimate the time and effort for the data extraction, transformation, and loading (ETL) functions.
Select the architecture that is just right for your environment; data warehousing is not a one-size-fits-all proposition.
Architecture first, technology next, and only then, tools.
Determine a clear training strategy.
Be wary of "analysis paralysis."
Begin deployment with a suitable and visible pilot to deliver early benefits.
Do not neglect the importance of scalability. Plan for growth and evolution.
Focus on queries, not transactions. The data warehouse is query-centric, not transaction-oriented.
Clearly define and manage data ownership ...