Automotive HMI development was historically feature and technology-driven. Over time, we witnessed a shift in focus from physical to cognitive issues, especially due to technology evolution and embedded HMI complexification. This made adapting automotive HMI development process a necessity to address human factors and cognitive ergonomics challenges in design/evaluation phases. It is in this context that car manufacturers enhanced the traditional systems engineering logic (V like model) thanks to the User-Centered Design cycle (UCD). But, despite this user centric approach, some customers' complaints and usability issues concerning automotive HMI are reported. Why is it so? To answer this question a research is underway. In this article, we (1) describe the work that led us to consider the research-practice gap as a candidate factor explaining why the user centric approach fails and (2) describe what we are planning to do as next steps.