In 2013, an undercover investigation by the BUAV raised serious concerns about the use, treatment and care of laboratory animals involved in regulated procedures at Imperial College, London. This led to an inquiry, set up by the college, which found deficiencies in the local ethical review process and a general lack of focus on the implementation of the Three Rs (Replacement, Refinement and Reduction). The Three Rs concept is the foundation of UK and EU legislation, but surveys of the published literature show that lack of its adoption is widespread. In spite of numerous guidelines, publications and publicity material extolling the benefits of the Three Rs to both animals and science, as well as substantial advances in the development, validation, and deployment of mechanistically-based non-animal methods, many scientists prefer to use traditional animal-based approaches. In addition, such scientists tend to pay less attention than they should to strategic planning, experimental design and the choice of appropriate statistical procedures. They are often unaware of the existence of replacement test methods to address all or some of their objectives, and are reluctant to develop and use new replacement methods. We explore some possible reasons for these shortcomings. We summarise the welfare and scientific effects of each of the Three Rs, and argue that: a) there is an urgent need for evidence to be made readily accessible to prospective licensees, which directly demonstrates the beneficial effects on animal welfare of the implementation of the Three Rs, separately and in combination, and the direct link this has with the quality of the scientific data obtained; b) a detailed systematic review of this evidence should be undertaken to augment the inadequate content of the prescribed Module 5 licensee training offered currently in the UK; c) such training (including that suggested in new EU-wide proposals) should be much more comprehensive, with stronger emphasis on the Three Rs, all parts of the syllabus should be fully examined, and there should be no exemptions from Module 5 training; and d) as the responsible Government department in the UK, the Home Office should take measures to tighten up its guidance for local ethical review, and its system of inspection of designated establishments, to obviate the justification for future undercover investigations.