Link traversal based query execution is a new query execution paradigm for the Web of Data. This approach allows the execution engine to discover potentially relevant data during the query execution and, thus, enables users to tap the full potential of the Web. In earlier work we propose to implement the idea of link traversal based query execution using a synchronous pipeline of iterators. While this idea allows for an easy and efficient implementation, it introduces restrictions that cause less comprehensive result sets. In this paper we address this limitation. We analyze the restrictions and discuss how the evaluation order of a query may affect result set size and query execution costs. To identify a suitable order, we propose a heuristic for our scenario where no a-priory information about relevant data sources is present. We evaluate the effectiveness of this heuristic by executing real-world queries over the Web of Data.