Abstract Aims The recently introduced technique of direct transseptal pacing of the left bundle branch is poorly characterized with many questions with regard to the optimal implantation strategy and safety concerns largely left unanswered. We developed a cadaver model for deep septal lead deployment in order to investigate the depth of penetration in relation to lead behaviour, lead tip position, and the number of rotations. Methods and results Five fresh human hearts and five lumenless, 4.1-Fr pacing leads were used for deep septal deployment simulations. The leads were positioned with the use of a dedicated delivery sheath and screwed into the interventricular septum at several sites progressively more distal from the atrioventricular ring with a predetermined number of lead rotations. During each lead deployment, the depth of tip penetration was measured and the lead behaviour was noted. Four distinct lead behaviours were observed: (i) helix only penetration, no matter how many rotations were performed, due to the ‘endocardial entanglement effect’ (43.1% cases) or (ii) ‘endocardial barrier effect’ (19.6% cases), (iii) shallow/moderate penetration, with ensuing ‘drill effect’ when more rotations were added (9.8% cases), and (iv) deep progressive penetration with each additional rotation, occurring when the ‘screwdriver effect’ was present (27.4% cases, including three septal perforations). These different lead behaviours seemed to be determined by the lead position—mainly the strength of the initial endocardial layer—and the number of fully transmitted rotations. Conclusion New insights into deep septal lead deployment technique were gained with regard to safe and successful implantation.