I'm no guru, but here is my 2 cents. Define what you mean in your context of a data mart.

If you are referring to a relational database using a star or snowflake schema within SQL Server, then you are still running into the same issues SSAS aims to minimize or eliminate. That's because SSAS pre-calculates / pre-aggregates stores of data in a format to make reporting much more efficient than the relational data mart within your SQL Server environment.

If you are having trouble understanding the end users needs or those needs are always changing where using and managing that cube is more of a pain than what it's worth, then that data mart optimized for reporting (reporting data mart as I call them) as much as you can be may be a worthy trade-off from cube performance to get your users to where they need to be.