because querying and aggregating your data with arbitrary conditions becomes much easier. Using SQL you describe what result you want and let the database figure out how to retrieve it.
In many document stores, you have to write manual map/reduce jobs telling the system how to get the data you want.
Most software development projects tend to be falling into two camps these days. Either they are simple and just need basic filtering/grouping in which case you don't need map/reduce. Or they are doing more advanced querying of data in which case Hadoop/Mahout is a better option.
So again why do I need a second data model that does nothing but get in my way ?