The overarching goal of this study is to perform a comprehensive bank stability analysis that is phenomenologically sound by considering both mass failure and fluvial erosion. The nature of this study is twofold. First, field and experimental analyses are conducted to generate data for channel cross-section properties, soil index properties, and mechanical and erosional strengths at two sites in a representative, midsize, midwestern stream in southeastern Iowa that is subjected to frequent flash floods and characterized by active fluvial erosion and cantilever failure. Second, the channel surveys and data obtained from the field and laboratory analyses are used as input parameters for an established one-dimensional, channel evolution model, namely, the conservational channel evolution and pollutant transport system (CONCEPTS, version 2.0, Langendoen and Alonso 2008), to estimate the factor of safety for mass failure (FSm) and fluvial erosion (FSf) and simulate the bank retreat as a result of either fluvial erosion or mass failure or the interaction between the two modes of erosion. In CONCEPTS, a bank profile can be divided into several layers, allowing the user to account for heterogeneity in soil properties. The results show that estimation of FSm must be complemented with the estimation of FSf for not underestimating mass failure. Otherwise, based on mass failure criteria alone, the stability analysis fails to consider the potential for the interconnection between bank toe undercutting and planar failure and may lead to the underestimation of mass failure over time. Second, bank soil heterogeneity plays an important role in bank stability analysis. The variability of mechanical and erosional strengths shown in this study, along the downslope of the banks, highlight the need to acquire both mechanical and erosional strengths for the three layers along a bank profile (crest, midbank, and toe) to improve the commonly adopted protocols that typically assume homogeneous, well-compacted soils along a bank profile. The predicted bank profile obtained from the model compares favorably with cross-sectional measurements obtained for a period of two hydrological cycles.