<p>An appropriate understanding of existing buildings is essential for assessing structural capacity and design of repair/retrofit solutions, although often only limited/incomplete information are available to analysts. This is particularly problematic when using advanced numerical tools based on discontinuum analysis – often required for out-of-plane (OOP) failure analysis of complex unreinforced masonry (URM) buildings. Discontinuum models simulate unit separation, cracking and collapse modes, advantageous for OOP failure modes, albeit requiring comprehensive inputs for building construction/connections to function properly. The prohibitive computational burden at larger scales seldom allows for treating uncertainties stochastically, thus deterministic nonlinear static analyses introducing varying degrees of simplifications are preferred. However, impact on structural engineering predictions remains mostly unknown. This study addresses this gap by quantifying the effects of common modelling assumptions on the OOP pushover response of a typical 19th-century URM stone church façade located in Montréal, Canada, using a Distinct Element Method (DEM) framework. Specifically, the numerical investigation examines the influence of façade discretization, boundary conditions and mechanical properties assigned to block interfaces while transitioning from meso- to macro-modelling scales. The results show that simplified masonry representation may be sufficient for force-based assessments when boundary conditions are accurately characterized. The results also highlight that introducing a compressive strength limit at block interfaces leads to reductions of up to 32% in predicted peak force and 48% in ultimate displacement. This work provides a quantitative basis for selecting appropriate modelling simplifications in structural assessment of heritage URM façades, helping balance computational feasibility and predictive accuracy.</p>