When the primary workload is analytics, reporting and large data scans, then a columnar db can be a good fit. The system can only read the columns specified in a query, rather than all of the columns in a record. This can minimize unnecessary data reads, and enhance the performance of aggregations and filtering. Another good reason to use column-based systems with compression is that similar data is stored clustered together. In evaluating database options, Datalance can assist you in considering query patterns, volume, performance requirements and the entire analytics architecture.