Agreeing and sharing common data definitions is very hard
We realize the value of having company- and even industry-wide domain models and taxonomies for automated processing of our data. But it is so difficult to achieve common consensus and completeness of those schemata. Everyone has their own slightly different view about our business, domain and world in general. With current methods we are forced to land on a least common denominator and more often than not nobody is completely happy. We waste time in meetings and discussions. Then when we need to fit the actual data to confirm the compromised model, it never fits perfectly.