The elaboration benefit on long-term memory (LTM) is well established and widely replicated (e.g., Galli, 2014). Yet, studies investigating the mechanisms underlying the effect are scarce. In two experiments, we tested two opposing hypotheses: (a) Elaboration has been assumed to benefit LTM by leading to meaningful integration of stimuli into existing networks of knowledge, and thereby, establishing potential retrieval cues (richness account, e.g., Craik & Tulving, 1975), and (b) that elaboration makes memory representations distinct from others, thus, facilitating memory search (distinctiveness account, e.g., Gallo et al., 2008). We manipulated combinations of elaborated features in order to create (a) rich and detailed or (b) distinct memory traces. Neither detailed nor distinctly elaborated words were remembered better than words elaborated with only one feature, which can be taken as direct evidence against the distinctiveness account. Although we also do not find evidence supporting the richness account, a slightly modified version of it could accommodate our results: Therein, elaboration benefits LTM retention by activating well-established semantic networks that have the potential to act as strong retrieval cues later on.