Abstract. The application of lipids in soils as molecular proxies, also often referred to as biomarkers, has dramatically increased in the last decades. Applications range from inferring changes in past vegetation composition, climate, and/or human presence to unraveling the input and turnover of soil organic matter (SOM). The molecules used are extractable and non-extractable lipids, including ester-bound lipids. In addition, the carbon or hydrogen isotopic composition of such molecules is used. While holding great promise, the application of soil lipids as molecular proxies comes with several constraining factors, the most important of which are (i) variability in the molecular composition of plant-derived organic matter both internally and between individual plants, (ii) variability in (the relative contribution of) input pathways into the soil, and (iii) the transformation and/or (selective) degradation of (some of) the molecules once present in the soil. Unfortunately, the information about such constraining factors and their impact on the applicability of molecular proxies is fragmented and scattered. The purpose of this study is to provide a critical review of the current state of knowledge with respect to the applicability of molecular proxies in soil science, specifically focusing on the factors constraining such applicability. Variability in genetic, ontogenetic, and environmental factors influences plant n-alkane patterns in such a way that no unique compounds or specific molecular proxies pointing to, for example, plant community differences or environmental influences, exist. Other components, such as n-alcohols, n-fatty acids, and cutin-and suberin-derived monomers, have received far less attention in this respect. Furthermore, there is a high diversity of input pathways offering both opportunities and limitations for the use of molecular proxies at the same time. New modeling approaches might offer a possibility to unravel such mixed input signals. Finally, the transformation and turnover of SOM offer opportunities when tracing such processes is the purpose of applying a molecular proxy while imposing limitations when they obliterate the molecular proxy signals linked to other phenomena. For n-alkanes several modeling approaches have recently been developed to compensate for (selective) degradation. Still, such techniques are in their infancy and information about their applicability to classes of components other than n-alkanes is lacking. All constraining factors considered can have a significant influence on the applicability of molecular proxies in soil science. The degree of influence strongly depends on the type of molecular proxy and the environmental context in which it is applied. However, the potential impact of the constraining factors should always explicitly be addressed whenever molecular proxies are applied in a soil scientific context. More importantly, there is still a serious lack of available information, in particular for compound classes other than the n-alkanes. Therefore, we...