The hypothesis that preparation before storage can increase the performance of a community of agents when adapting and reusing knowledge will be explored. Community performance is measured by how many time steps it takes to solve a problem. Currently, the system can be evaluated by the efficiency of the traces stored in collective memory, where the efficiency of a trace can be defined to be the performance it would yield upon reuse for the exact same problem. Although this is perhaps not as meaningful as the average expected performance in all situations where the trace could be used since it ignores adaptation costs, it is tractable and provides a reasonable basis for evaluation. When the system is more mature, an empirical evaluation of the merits of data preparation can be determined by tracking the performance of the community through a suite of test problems. No comparisons to other methods are planned at this time. We would eventually like to include both planner and memory costs in a more global evaluation of system performance. A significant contribution of our research will be highlighting the opportunity for, and benefits of, preparing knowledge before adaptation and reuse.