Résumé : Many studies have already been published concerning autogenous shrinkage in cementitious materials. Still, no consensus can be found in the literature regarding the determination of the time-zero to initiate the recording of autogenous shrinkage. With internal curing agents, a correct evaluation of their efficiency depends on an appropriate choice of the time-zero. This study investigates different approaches to estimate the time-zero for cement paste mixtures with and without superabsorbent polymers as internal curing agents. The initial and final setting times were determined by an electronic Vicat and ultrasonic pulse velocity measurements (UPV); the transition point between the fluid and solid state was determined from the autogenous strain curve; the development of the capillary pressure was also studied. The choice of time-zero before the transition point led to higher values of shrinkage strain that should not be taken into account for autogenous shrinkage. A negligible difference was found between the strains when the final setting time and the transition point were taken as time-zero. Considering the artefacts and practical issues involving the different methods, the use of the transition point from the autogenous strain curve is the most suitable technique for determining the time-zero.