The Kantian imperative to treat persons with respect by never using them as mere means and always as ends-in-themselves has a wide currency in contemporary ethical thinking. But how exactly do we use a person, whether ourselves or others, as a mere means? While there are a number of competing views which try to answer this important question, each of these views has, I shall argue, important intuitive problems. In this paper I shall seek to overcome these problems by defending an intuitively appealing account of what constitutes using ourselves and others as mere means. Roughly, we use ourselves as a mere means if we treat our capacity to make choices as less important than any other end, and we use others as mere means if we fail to obtain their possible consent when we interact with them. Finally, we shall also look briefly at how using others as mere means differs from failing to treat them as ends-in-themselves.