In this paper, we introduce the concept of "phonetic compliance," which is defined as the intrinsic individual ability to produce speech sounds that are unusual in the native language, and constitutes a part of the ability to acquire L2 phonetics and phonology. We argue that phonetic compliance represents a systematic source of variance that needs to be accounted for if one wants to improve the control over the independent variables manipulated in SLA experimental studies. We then present the results of a two-fold proof-of-concept study aimed at testing the feasibility of assessing phonetic compliance in terms of gradient. In study 1, a pilot data collection paradigm is implemented on an occasional sample of 10 native French speakers engaged in two reproduction tasks involving respectively vowels and aspirated stops, and data are analyzed using descriptive statistics. In study 2, complementary data including L1-typical realizations are collected, resulting in the development of a first set of indicators that may be useful to appropriately assess, and further refine the concept of, phonetic compliance. Based on a critical analysis of the contributions and limitations of the proof-of-concept study, general discussion formulates the guidelines for the following stages of development of a reliable and valid test of phonetic compliance.