In recent years, biodiversity offsetting has been adopted as a conservation strategy in many different countries. Biodiversity offsets are mechanisms used to compensate for ecological impacts resulting from development projects, especially those on non-built land uses. They usually rely on an equivalence principle based on achieving strict equality between the area that is developed and that which is offset. This approach still remains very controversial. This article explores an alternative offset design method, where we look at biodiversity from a functional perspective and conceive public policies that aim to conserve biodiversity by maintaining important structural features of the landscape, not limited to the proportion occupied by each land-use. We develop a spatially explicit land use change model to implement our geometric-based compensation method and we define three different versions of the public compensation policy. We apply the model to real case studies in two French municipalities and we compare the cost and feasibility of compensation under the different public policies. We find that offsetting is easier and cheaper when public policies aim to conserve no more than the targeted semi-natural land area, but this approach has major ecological limits. When considering more complex geometric properties of the landscape (and therefore higher ecological expectations), compensation becomes, on average, more difficult and more expensive. Our work shows how new approaches to ecological compensation could be defined and how models could help select the best options in the field.