The concept of similarity is central to marketing research. Its fields of application are very varied. Yet this concept, which is more complex than it seems, is still poorly understood. It must be distinguished from other closely related concepts such as similitude, difference, fit and congruence. This paper shows that the type of similarity used (literal or relational) has different effects on consumers' cognitive processes. This can dramatically alter the results of an experiment. The same applies to the type of measurement used.