Habitat loss is a primary threat to biodiversity across the planet, yet contentious debate has ensued on the importance of habitat fragmentation ‘per se’ (i.e., altered spatial configuration of habitat for a given amount of habitat loss). Based on a review of landscape-scale investigations, Fahrig (2017; Ecological responses to habitat fragmentation per se. Annual Review of Ecology, Evolution, and Systematics 48:1-23) reports that biodiversity responses to habitat fragmentation ‘per se’ are more often positive rather than negative and concludes that the widespread belief in negative fragmentation effects is a ‘zombie idea’. We show that Fahrig's conclusions are drawn from a narrow and potentially biased subset of available evidence, which ignore much of the observational, experimental and theoretical evidence for negative effects of altered habitat configuration. We therefore argue that Fahrig's conclusions should be interpreted cautiously as they could be misconstrued by policy makers and managers, and we provide six arguments why they should not be applied in conservation decision-making. Reconciling the scientific disagreement, and informing conservation more effectively, will require research that goes beyond statistical and correlative approaches. This includes a more prudent use of data and conceptual models that appropriately partition direct vs indirect influences of habitat loss and altered spatial configuration, and more clearly discriminate the mechanisms underpinning any changes. Incorporating these issues will deliver greater mechanistic understanding and more predictive power to address the conservation issues arising from habitat loss and fragmentation.