Building on available theory, this work uses controlled laboratory experiments to investigate the budgetary and the economic performance of competitive tenders for allocating conservation contracts to landholders. Experiments have been replicated in two different countries to check for robustness of results. We find that auctions outperform the more traditional fixed-price programs only in the one-shot setting. With repetition, the auctions quickly lose their edge. The budget-constrained auction performs similarly to the target-constrained in the one-shot setting but appears more robust to repetition. Our results suggest that previous estimates of conservation auction performance are too optimistic, and we propose a method for improving such estimates.