Indispensability arguments for mathematical realism rely on indispensability claims about mathematics that deserve critical examination. In the "Varieties of dispensability arguments" Panza and Sereni review some major indispensability claim relative to different notions of descriptive or explanatory indispensability and find them wanting. In this paper we suggest a new standard of indispensability which seems to be more in line with the spirit of deference to science as regards normative matters that inspired the indispensability arguments in the first place. Because dispensability/indispensability claims are to be assessed as least in part relative to epistemic costs one may or may not find acceptable when dispensing with the mathematical apparatus, indispensability claim in fact hinges on disputable normative considerations. We submit that the final judge on matters of indispensability - whether descriptive or explanatory - of mathematics to science should be the scientists themselves, thus taking one more naturalistic step by naturalizing indispensability claim themselves.