Over the years, we have found that some real-world phenomena are not as clear to regular people as what Wikidata could store. A common example is a museum, where the building it is in often is conflated with the organization (often the legal entity) operating the museum. Sometimes this makes sense, but sometimes it could also lead to odd property combinations, for example that something both has the property "floors above ground" and "total revenue". Another example is a journal publisher with only one journal. Some users have argued that there is no value to creating one of each of these items, whereas others argue that this is needed to keep the data model clean and the results from queries predictable. In this session, we will discuss a number of common instances of this to see what advantages and disadvantages each of these ways of modeling has, and also try to draw some general conclusions that could be applicable to other similar occurrences.