Market segmentation allows you to focus your scarce marketing resources and appeal to potential customers in ways that are most likely to get them to become loyal customers. Using segmentation, you can speak to the needs and interests of different groups; you can determine whether there is a product/service fit in high opportunity.
At one of my client place the Customer Base was around 2 millions approximately, and there were around 1500 Marketing Attributes on which the Marketing Department of my client profiles the customer to focus their marketing campaigns. Because of this large Customer Base and large number of Marketing Attributes when marketing department profile their customers they were waiting for large time (For Ex: to profile customer based country = US and City = California) their wait `time to see the profile set was approximately around 30 min. To avoid the long wait time the below solution was proposed.
Use fast find to create a buffer for storing the business partner data that you often use with Segment Builder profile modeling. Fast find reduces the frequency of database accesses to improve search efficiency in large databases.
Fast find functionality using Classical Segmentation accelerates profile modeling in Segment Builder. It provides fast access, and a function for counting and reducing the number of database accesses. Primarily, you use it when you need to search millions of data records for certain attributes. When you define a profile model in Segment Builder, you can monitor how many business partners meet the profile criteria with attribute values. If you access large databases, it can take a long time to determine the attribute distribution. In this case, you should use fast find because it can improve search efficiency. During data retrieval, the system compiles the segment’s business partner data along with corresponding attributes into a searchable index or table within the Text Retrieval and Information Extraction (TREX) cache.
Detailed steps on how to implement Fast Find were given in my Wiki