This case is a detailed account of the planning and execution of the data collection for the Unions in Context study, which is a probability-based survey among lesbian, gay, and heterosexual couples and families in the Netherlands. On one hand, we situate the Unions in Context study in the context of the status quo of the field. We discuss the opportunities and challenges of the different types of pre-existing data sources that allow for the study of same-sex couples and families, such as convenience sample surveys, register and census data, and large-scale social surveys. In this discussion, we highlight the contribution of the Unions in Context data to the field. On the other hand, we outline the practical considerations made during the planning phase; the most commonly faced challenges when studying lesbians, gay men, and bisexuals quantitatively; and the solutions we thought of to address them. Specifically, we explain the steps of how we have succeeded in designing a sampling frame based on population registers to draw a random sample of same-sex and mixed-sex couples. This step-by-step account should enable other researchers interested in studying this difficult-to-reach population to replicate our approach.