The main problem of research-driven design is determining granularity of research. Monitoring brain activity of museum visitors is too much, asking nothing is too little. To spend resources well, a venue should know what kind of research contributes to which design outcomes.