In order to support teachers to engage themselves in a learning process to improve their techno-pedagogical skills, and to relieve instructional designers (IDs) of their workload, we have previously designed a behavioral computer model of teachers that allows their evaluation from a LMS mastery perspective, and paves the way to teaching analytics and recommendations. To instrument this model and provide teachers and IDs with a digital support that best meets their expectations, we present here the user-centered method we followed to (i) evaluate our model to their perceptions, (ii) determine consistent uses relative to their needs, and (iii) develop iTeachApp, an application that provides teachers with a personal analytics dashboard and automatic recommendations, and offer IDs institutional analytics, to detect specific profiles of teachers and get insights of common LMS behaviors at different scales.