This paper explores the efficiency of multi-tenancy in data warehouse environments by evaluating three distinct architectural approaches: Separate Schemas, a Single Schema with Tenant Identification, and Tenant Grouping. Through comprehensive performance experiments simulating realistic multi-tenant scenarios, we assess key metrics such as throughput, query execution time, and data isolation capabilities. Our findings reveal trade-offs between complete data isolation and shared schema approaches, highlighting the influence of resource management and hardware provisioning on performance. Additionally, Tenant Grouping emerges as a promising optimization strategy, balancing data segregation with performance efficiency. The study provides practical guidance for selecting multi-tenancy architectures in cloud-based data warehouse systems, considering factors such as scalability, performance isolation, and resource constraints.