Relational OLAP tools and other database applications generate sequences of SQL statements that are sent to the database server as result of a single information request provided by a user. Unfortunately, these sequences cannot be processed efficiently by current database systems because they typically optimize and process each statement in isolation. We propose a practical approach for this optimization problem, called ``coarse-grained optimization,'' complementing the conventional query optimization phase. This new approach exploits the fact that statements of a sequence are correlated since they belong to the same information request. A lightweight heuristic optimizer modifies a given statement sequence using a small set of rewrite rules. Since the optimizer is part of a separate system layer, it is independent of but can be tuned to a specific underlying database system. We discuss implementation details and demonstrate that our approach leads to significant performance improvements.