Microsoft Purview Data Loss Prevention

Microsoft Purview DLP looks very much like a solution still in development, with lots of basic functionality you would be expecting missing (e.g. simple use case - blocking upload of any file to a certain site cannot be achieved), some that are half delivered (e.g. blocking of applications is done completely on any data they access, instead of blocking just outbound traffic) and some features that are indeed in place. Microsoft Purview DLP doesn't seem to have been consistently designed with sound principles being pushed consistently across the board, rather it seems to have been slapped together from pieces, where those pieces don't look or work the same (e.g. when it comes to scoping policies to large organizations, scope might be dynamic and the tool simply doesn't support dynamic groups across all its workloads). Sometimes you can come across an issue that makes no sense at all (e.g. you have a OneDrive policy that blocks access to files and then it generates absolutely no alerts) and Microsoft loves to give answers on everything that's not there or plain out wrong that "it's by design". This highlights the fact that on top of Purview DLP's current development status, Microsoft has adopted a support model that is neither flexible nor timely, even in critical cases were the customer is articulating significant business impact (e.g. if you get a modification you require accepted by the Microsoft DLP product group, it can sit in a development hell limbo with no predicted release date for it, even for years). Last but not least, Microsoft's official documentation on the product is also not always up-to-date or detailed enough to help prepare you on all angles that you need to consider before implementing something. The overall experience is that this is a tool that's not functioning on good principles that are pushed across the board, but rather by exceptions, hence the learning curve is quite steep and you should always brace yourself that upon the next use case you're trying to implement there is a good chance you'll hit a "catch" that will ruin your best planning or most logical assumptions.