Data warehouses combine data from various, often heterogeneous sources. The hetero-homogeneous approach to data warehouse modeling allows for the representation of heterogeneities while preserving the advantages of homogeneous models. A hetero-homogeneous model allows for the stepwise, controlled introduction of additional levels and measures for different sub-hierarchies and sub-cubes of an otherwise homogeneous model. The talk gives an introduction to the conceptual modeling of hetero-homogeneous data warehouses using multilevel objects. The talk further presents a proof-of-concept implementation for the management of hetero-homogeneous data warehouses, inclduing a graphical OLAP client as well as a data definition and query language for hetero-homogeneous data.