Data-driven learning (DDL) is a pedagogical approach that assumes the use of large amounts of authentic language data to facilitate foreign language teaching and learning. The approach involves students discovering linguistic phenomena through data analysis, acting as researchers and developing their ability to observe and infer. There are two main approaches: inductive and deductive DDL. Inductive DDL is based on the observation of data in order to reach the formulation and generalisation of a rule, whereas deductive DDL centres on the statement of a rule, which is then observed within a corpus. The inductive approach encourages student autonomy and develops their curiosity, acting on their motivation; the teacher is an ‘instructor’ (Smart, ReCALL 9:5–14, 1997, p. 186) who simply acts if necessary. The deductive approach, on the other hand, is more structured and guided by the teacher, who keeps their role as the holder of knowledge; the student observes the application of the rule stated in the materials provided by the teacher. The choice of one approach over the other depends on numerous factors including the educational context, learning objectives, students’ skills and their learning styles.