Efficient and effective lifelong learning requires that learners can make well informed choices from a vast amount of learning opportunities. This article proposes to support learners by drawing on principles of self-organization and indirect social navigation; by analysing choices made by learners who went before and feeding this information back as advice to present learners. A tool, developed to offer indirect social navigation through collaborative filtering, and tested in a controlled experiment, showed positive effects on effectiveness (progress and completion rates) though not on efficiency (time taken to complete and learners' perception of learning track efficiency). Learners in the experimental group were not faster at completing all available units of learning nor did they evaluate learning track efficiency more positively. Responses on the five-point scale measuring learning track efficiency seemed to indicate a dichotomy rather than an ordinal scale. Learners stated either that studying the units of learning in a different order might have led to increased understanding of the content or that a different order would not have led to greater efficiency. Analysing this dichotomy in relation to learning track characteristics by means of the decision tree induction data mining technique, revealed that a small subset of navigational choices discriminated between a positive evaluation on the one hand and doubts regarding the efficiency of the learning track on the other hand. These results suggest that positive effects for efficiency might be achieved, by incorporating learners' overall evaluations of learning tracks into the advice. (Contains 1 table and 3 figures.)