Language teachers find themselves constantly challenged by the different types of learners in their classrooms. At the same time, researchers in second language acquisition (SLA) struggle with the problem of how to account for learner differences in SLA models and theories. This paper examines learner differences from the perspective of both teacher and researcher by comparing the first few hours of guided SLA of two learners with quasi-identical profiles exposed to identical target language input. The data reported here were collected during a Polish course designed for French students who had had no previous contact with Polish. We controlled the input provided to our learners in its entirety from the moment of first exposure. This allowed for a quantitative comparison of our learners' performance in Polish on a given task relative to the input received and to each other. Grammatical phenomena studied were primary constituent word order, the placement of the negative particle nie, and verbal morphology. Most SLA models would predict a similar acquisitional path for our two learners who share the same first and second languages, and are of the same age and gender. Contrary to predictions, our learners approached their first hours of Polish quite differently. We will discuss the results in terms of their potential contribution to classroom teaching and to the formulation and elaboration of models and theories in SLA, emphasizing the importance of continued controlled empirical data which can and should contribute to our understanding of individual variability.