Relational databases – once highly prized for their querying capabilities and transaction management – are now being supersede by a new paradigm, NoSQL (Not Only SQL), that targets their shortcomings of scalability and big data performance issues. The most popular NoSQL representative is MongoDB, an open-source project with strong enterprise support. TPC-H (Transaction Processing Performance Council- Ad-hoc), the de facto decision support benchmark system for business analytics, was used to determine the viability of MongoDB. The TPC-H schema was mapped to the MongoDB schema-less representation and the 22 TPC-H queries are written in the MongoDB query language. Several configurations are investigated, both for schema representation as well as for the queries. For schema representation the configurations that were considered were a fully denormalized (embedded) model and a fully normalized (referenced) model. MongoDB was compared against SQLite, the most popular embedded relational database. The results reveal that using the denormalized model for mongoDB yielded greater flexibility in querying because the queries were ran against a single collection. MongoDB performed effectively with moderate amount of data, but as documents in the collection increased in size complications will arise. In conclusion, MongoDB was able to compete with SQLite on complex queries though ins some cases MongoDB’s embedded model traversal latency was larger compare to SQLite. Overall MongoDB proved itself to be a viable data base, its scalability and performance rivaled that of SQLite and in some cases surpassed it.