Absolutely, that's a very practical approach to take. In reality, the optimal solution for databases doesn't necessarily match what people would select right no

Absolutely, that's a very practical approach to take. In reality, the optimal solution for databases doesn't necessarily match what people would select right now from scratch. If a system is working fine, the expense and risk of moving may end up being far greater than any optimization efforts on the existing database. This is also why I believe profiling tools, explain plans, and indexes are all useful; there are plenty of cases where MongoDB is sufficient but still needs optimization...