Many organizations now purchase and customize software rather than build information systems. In this light, some argue that high-level data modeling no longer has a role. In this paper, we examine the contemporary relevance of high-level data modeling. We addressed this issue by asking 21 experienced data-modeling practitioners to reflect on their work and to give their opinions on trends and future directions in high-level data modeling. We analyzed transcripts of our interviews with them using Klein and Myers’s (1999) framework for qualitative research. We found considerable variation in the practice of high-level data modeling. We also found that high-level data modeling is still considered important, even though organizations ultimately may purchase off-the-shelf software. The reason is that high-level data modeling assists organizations to obtain clarity about IT project scope and requirements, thereby reducing the risk that costly implementation mistakes will be made.