This paper empirically measures the prevalence of heterogeneous technologies in a sample of small-scale agricultural producers as an answer to structural conditions and market risks. Such risks are closely linked to the effects of land fragmentation and the degree of market integration. We use the empirical case of Kosovo as a transition country to investigate the efficiency effects of land fragmentation by simultaneously considering the effects of market integration. Different to previous studies, we assume that land fragmentation and market integration lead to the prevalence of heterogeneous technologies allowing farm households to respond more efficiently to exogenous price and policy shocks given their fragmentation and subsistence situation. The empirical work links the latent class frontier method to the estimation of a directional output distance function. We estimate beside primal technology measures also dual Morishima type elasticities of substitution investigating changes in production decisions based on relative shadow price changes.