The data cleanup processing requirements will vary from data warehouse to data warehouse. There may be no data cleanup, or there could be a major match and cross-reference operation to perform, with each record being verified against othel data. For example, in a call record data warehouse you may need to match each call record with the customer information to get customer ids onto the record. This whole process is again a serialization, because the index and aggregation builds probably cannot proceed until any data cleanup has completed.