Using the landmark transformation technique, previous research on multiple landmark use by pigeons (Columba livia) has shown that pigeons tend to encode a goal location relative to a single landmark instead of a group of landmarks. Based on these findings, it has been assumed that pigeons are unable to encode a goal location relative to the entire configuration of a landmark arrangement. The current experiments sought to determine whether making the configuration of landmarks in an arena a discriminative cue to the location of food would result in accurate piloting to the goal, and whether the goal location would be encoded relative to the configuration of landmarks, instead of just one. Four uniquely colored and shaped landmarks were presented in one of two configurations within a square-shaped arena. The buried food goal was located in different areas of the arena based on the landmark configuration. In Experiment 1, pigeons were given unbaited search trials with each of two configurations, as well as trials when the landmarks were not present in the arena. Pigeons searched accurately with the two configurations, but searched between the two goal locations on trials with no landmarks present. In Experiment 2, after training with the two configurations, pigeons were given unbaited tests on which the arena was rotated and the configurations were approached from novel perspectives. Results showed disruption in search accuracy on novel perspective trials. In Experiment 3, pigeons were give experience with two approach directions for each configuration in training and were tested with novel perspectives. There was no improvement in search accuracy after such training. In Experiment 4, test trials were conducted where one landmark was shifted away from its position in training to determine whether a single landmark was used to encode the goal (by resulting in a shift of pigeons' searching). Pigeons did not shift their searches to follow any specific landmark, indicating that information from the remaining landmarks was sufficient to search accurately. These experiments demonstrate that landmark piloting in pigeons is more flexible than previously thought, and that procedural differences may be the cause of discrepant findings.