In this paper, the orderbatching problem in warehouses is investigated. Batching, or clustering orders together in the picking process to form a single picking route, reduces travel time and, as such, this phenomenon can be encountered in many warehouses. The reason for its importance is that orderpicking is a labour intensive process and, by using good batching methods, substantial savings can be obtained. The batching and routing problems are complex to solve. In practice, simple methods are used for the batching problem, such as first-come first-served (FCFS), i.e. combine orders in the route in the arriving sequence until the pick-device is full. Once clusters of orders have been formed, the calculation of the travel time for the routes requires the solution of a number of travelling salesman problems (one for each route). Two groups of heuristic algorithms are evaluated: the Seed algorithms and the somewhat more complex (and CPU time consuming) Time Savings algorithms. The performance of the algorithms is evaluated using two different routing strategies: the so-called S-shape and Largest gap strategies, which are well known in theory and practice. The heuristics are compared for travel time, number of batches formed and also for robustness. An algorithm is robust if it can be applied to any warehousing situation and still yield good results. Since the problem has to be implemented in existing Warehouse Management Systems and has to be solved online many times a day, the simplicity, transparency and CPU time are also important in most warehousing situations. It is demonstrated that even simple order batching methods lead to significant improvement compared to FCFS. Seed algorithms are best in conjunction with S-shape and a large capacity of the pick device. Time savings algorithms perform best in conjunction with Largest gap and small pickdevice capacity. If CPU time becomes important, then using simple Seed algorithms should be considered.