When you're looking for a column oriented database, you typically need to be doing large analytical queries, aggregations, reporting or business intelligence operations. It can rapidly scan specific fields within a large number of records, rather than reading the whole record. This can help to optimize performance and, in some cases, also lower storage consumption, due to better compression. But might not be suitable for workloads with frequent individual record updates or transactional activity. First, I'd look at how queries are used, the size of the data, how concurrent it is and how often it was updated. If you're looking for assistance with assessing an analytics architecture, you can try out Datalance.