Building a new data warehouse (or converting to a new data warehouse) starts with asking the right questions and finding the answers that provide a starting point.

From a project point of view start with the following questions.

  * Why has the project started?
  * Who is your client, what type of person is he or she and what role / place in the organisation does he or she have?
  * What will the budget be to build the infrastructure (hardware and software) and create your loading processes?
  * Are you working in a team and if so, what technologies are familiar or lie in the ambitions?
  * If you are working in a team, what is the current spirit of the team? What type of people are there?
  * What are the expectations of your client in terms of desires and time constraints?

From a technical point of view start with a quick few questions to get a general idea of what the architecture should be able to cope with.

  * Will the data be be delivered or do I need to export it myself?
  * What type of sources do I need to load data from? Databases, flat files, video / audio, etc.
  * Will my data warehouse need real time data, near real time data, daily/weekly/ monthly/etc batches? How probable is it that I will have to move to an architecture that supports near real time data or even real time data?
  * Will there be expert users querying the data warehouse directly or will I be creating data marts for them to query?
  * Do I need to prepare data for a reporting and/or analysis tool?
  * What type of hardware and software is already being used for other applications?
  * Who will approve your design for infrastructure?
  * Will you be administrating the infrastructure yourself?

Write this down somewhere so you can look back at it, preferably with a date and who gave the answers to your questions. Use it as a reminder or to evaluate when the first deliverables are being finished.