Biodiversity is a difficult and controversial thing to describe and measure. Yet managers of working forests have increasing incentive to establish baselines, measure performance, and track progress in biodiversity conservation. Forest certification is one such incentive, requiring a means to document and quantify biodiversity conservation practices across large forest ownerships. The need is for assessment tools that are practical to apply, credible to conservation biologists, operable at multiple scales and across broad regions, and compatible with forest management planning and inventory systems. We offer a conceptual proposal for an assessment approach that satisfies these requirements. Our approach relies on a coarse-filter ecological classification. It emphasizes landscape composition, without explicit measures of structural or functional integrity that can be difficult to develop and defend. The assessment is not limited to an inventory of "special areas" or sites of exceptional conservation value; we propose to assess the entire forest landscape including the impact of plantations and the dynamics of even-aged management. We include landscape-level measures of richness, conservation value, and representation. We discuss challenges to implementation and example applications.