Sometimes it's not for lack of trying, though. I've done a decent amount of data model design in my career and as I got more experienced, I learned to make a point of asking the business people what specific questions they'd need the data to answer. Very often I got only very vague answers that were of little to no use in informing my design. I've seen this repeatedly at a variety of companies over the years. The business people are often winging it as much as the engineers are.