While there may be general agreement on what it means for a formal method to be usable (e.g., ease of modelling, automated and scalable analysis), there is no consensus in the software-engineering or formal-methods communities on what strategies lead to more usable formalisms. In this paper, we aim to raise discussion around such strategies by proposing fourteen concrete recommendations for achieving practical formal methods. Our recommendations apply to research in formal modelling, automated analysis, and automated transformation (e.g., transforming a model into the input to an analysis tool). Our recommendations focus on formal methods for functional requirements of software product lines, as per our experience in this area as part of a research project in collaboration with an automotive manufacturer; however, most of the recommendations apply to formal methods in general. We also provide a brief overview of a formal modelling language and an underdevelopment tool chain that realizes our recommendations.