Understanding how and why cells cooperate to form multicellular organisms is a central aim of evolutionary biology. Multicellular groups can form through clonal development (where daughter cells stick to mother cells after division) or by aggregation (where cells aggregate to form groups). These different ways of forming groups directly affect relatedness between individual cells, which in turn can influence the degree of cooperation and conflict within the multicellular group. It is hard to study the evolution of multicellularity by focusing only on obligately multicellular organisms, like complex animals and plants, because the factors that favour multicellular cooperation cannot be disentangled, as cells cannot survive and reproduce independently. We support the use of Saccharomyces cerevisiae as an ideal model for studying the very first stages of the evolution of multicellularity. This is because it can form multicellular groups both clonally and through aggregation and uses a family of proteins called 'flocculins' that determine the way in which groups form, making it particularly amenable to laboratory experiments. We briefly review current knowledge about multicellularity in S. cerevisiae and then propose a framework for making predictions about the evolution of multicellular phenotypes in yeast based on social evolution theory. We finish by explaining how S. cerevisiae is a particularly useful experimental model for the analysis of open questions concerning multicellularity.