1. Operational Data Store has very low data latency. Data moved to ODS mostly on event based rather than time based ETL to Data Warehouse/Data Mart.
2. ODS is more closer to OLTP system. We don't normally prefer to store aggregated data in OLTP. So it is with ODS.
3. Unlike data warehouse where data is HISTORICAL, ODS is near real time(NRT). So data aggregation is less important is ODS as data keeps changing.