In this paper, we first argue that the word "innovator" may be inappropriate to describe early triers of new grocery products, because such products are often not real innovations. Besides, an innovator is classically defined as the adopter of a new product, and the definition of "adoption" poses problems for such grocery products. Most of the paper is devoted to the description of a series of empirical analyses. These analyses lead to the final conclusion that the observed overlap in early trial across product classes might be misleading. We show that persons who are early triers of new products in multiple product classes do not necessarily do so because of some innovative disposition toward trying new products, that would allow us to characterise them as "generalised innovators". Rather, most (or all) of this overlap in early trial could be due, spuriously, to the overlap in purchase frequency across product classes, and to the relationship, in each class, between early trial and purchase frequency.