To be able to predict the spatial dynamics of species, knowledge on the ability and frequency of individual dispersal events are a fundamental requirement. Recent advances in genetic techniques have enabled ecologists to estimate the effect of landscape features such as roads, rivers and unsuitable habitat on dispersal as the needed resources of such studies are decreasing steadily. Least-cost path analysis is the most common approach to study the effect of landscape features on population structure. The basic idea is simple – various landscape models represented by distance matrices that incorporate the effect of landscape features on dispersal are compared to genetic distance matrices. We present an approach that allows us to predict the likely performance of an intended study design by simply studying the geometric properties of the sampling design. We demonstrate the application of the approach employing a simulation study and a real-world example using a well studied possum meta-population. Finally we present the implementation of the approach within the R-package PopGenReport. We demonsrate that by using approach the sampling design of an intended study can be optimised, which leads to more precise estimates of the effect of landscape features on dispersal, a better abilityto predict the spatial dynamics of a species and ultimately it will allow us to conserve threatened populations more efficiently using landscape-genetic approaches.