This paper compares four sensor placement strategies which differ according to their evaluation criteria (EC). The first involves the minimisation of the expected number of candidate models, and the second is based on maximising joint entropy. The first methodology shows better results in terms of diagnostic performance. However, the second is promising due to faster execution time. The third strategy is a combination of the first two. Finally, a fourth strategy involves consideration of the cost of the sensor placement at each location in addition to the EC of the third strategy. The four strategies are evaluated in terms of performance, computational load and cost. Since there is only mild competition between the three criteria, a hierarchical multicriteria decision-making approach is employed to identify the best sensor placement strategy. Two case studies are used for illustration. The results show that the sensor placement strategies are useful for identifying optimised sensor configurations for new configurations as well as for evaluating the performance of existing sensor configurations. Using a hierarchical multicriteria decision-making technique, the fourth sensor placement strategy satisfies all criteria well, making it the best strategy.