Abstract Harm-benefit analyses (HBAs) are becoming de rigueur with some governmental regulatory agencies and popular with local institutional animal care and use committees (or their equivalents), the latter due, in part, to the adoption of HBAs as an international accreditation standard. Such analyses are employed as an attempt to balance potential or actual pain or distress imposed on laboratory animals against scientists’ justifications for those impositions. The outcomes of those analyses are then supposed to be included in an official assessment of whether a given animal protocol should be approved as proposed. While commendable in theory as a means to avoid or minimize animal suffering, HBAs come with a flawed premise. Establishing an accurate prediction of benefit, especially for so-called “basic” research (vs “applied” research, such as in vivo testing for product development or batch release), is often impossible given the uncertain nature of experimental outcomes and the eventual value of those results. That impossibility, in turn, risks disapproving a legitimate research proposal that might have yielded important new knowledge if it had been allowed to proceed. Separately, the anticipated harm to which the animal would be subjected should similarly be scrutinized with an aim to refine that harm regardless of purported benefits if the protocol is approved. The intentions of this essay are to reflect on the potential harm and benefit of the HBA itself, highlight how HBAs may be helpful in advancing refinements, and propose alternative approaches to both parts of the equation in the assessment process.