The paper deals with assessment of the influence of adaptive e-learning as a part of learning analytics on learning effectiveness of primary school pupils. E-learning exercises containing implemented adaptive elements were created in accordance with the Bloom's Taxonomy. Within the pilot study the authors detected high percentage success rate during e-learning exercise completion. This leads to formulation of the question „Can any e-learning exercise of lower cognitive levels of Bloom's taxonomy be skipped without affecting the cognitive thinking for solution of the e-learning exercises on upper cognitive levels of Bloom's taxonomy?” To answer the question, the algorithm of adaptive e-learning was defined and hypotheses were established. The research was carried out as pedagogical experiment comparing the results of both experimental and control groups of pupils. The research hypotheses were confirmed by statistical analysis of the research data. The results confirm that adaptive features of e-learning can be implemented in the primary education. The research results confirm the fact that educational objectives can be achieved with some pupils more effectively. Consequently, the implementation of adaptive elements into e-learning at the primary school supports an individual approach when completing e-learning exercises according to the principle of cognitive computing. •Adaptive e-learning (AE) application in primary school education were studied.•Algorithm for primary AE based on concept of learning analytics was proposed.•Adaptive algorithm of primary AE meets the typical features of cognitive computing.•Learning effectiveness (LE) was clearly and exactly defined.•It was confirmed that the LE of primary e-learning increases by utilization of AE.