When integrating systems, the focus is on creating the right container to store data that may come when the system goes online. Schemas, joins, views… in systems that were traditionally not built to store the kind of data you want. When the system comes online, you realize the requirements have already changed. There is still no data. Data updates: today most integrations are one-directional. e.g. environmental data -> CAD. Bi-directional integrations are costly since you need to customize 2 systems instead of one. So any feedback loops are typically lost in emails, handwritten notes and issue tracking systems. When connecting systems, you are essentially doing what every single person that’s done an integration, has done before – mapping master data. For every data source, you are doing this again. and again. Laborious, prone to errors and without the ability to continuously make corrections while controlling this process. As a result, only a handful of companies have multi-dimensional data.