Customer Orders, with Multiple Demography details

Newbie here. I have the following tables in 3NF

The business process identified is Recieve Customer Orders.

Customer(CustomerID...)
Demographics(DemographicID...)
and CustomerDemographics(CustomerID,DemographicID) -- A customer could belong to one or more demography.

Orders(OrderId, CustomerID)..

When I turn this into a dimensional model, how does one flatten the Customer-Demographics data, around the fact table for Orders.

One of the approaches I considered is to create a dimension for each demography and have the FKs in the Order fact table and create this for frequently used demography alone.

How does one solve this generically?

When an order is created, does it belong to any demographic? If not, you need to ask the business users why they want to report including the demographics. If they are going to build reports with the different orders by demographics, an order may be included in multiple reports if the many-to-many relationship is included in the fact table. Is this what they require?

Yes, indirectly. A customer is part of one or more demographic and an order is placed by the customer. The requirement is to analyze the order/order items by demographic. So an order may be included in multiple demographics.

This is generically refered to as multivalued dimensions. Google the term, there is probably something out there that explains how to implement them. Its kind of hard to explain how to do it within the confines of this forum.

Start with

http://intelligent-enterprise.informationweek.com/010810/412warehouse1_1.jhtml

Thank you for the link.