I work on a team of about 20 people. Our portfolio management data (schedules, key deliverables, etc) used to be spread out in different files that were e-mailed to each other. This is fairly small as far as data sets go, but can still quickly get out of sync as everyone here can probably imagine.

We've been able to improve on this by managing data into List data structures on our SharePoint site, and referring everyone there for creating/reading/etc. This gives us a single-source-of-truth, some version control/approval capabilities, lets us do some simple cross-referencing of our data, and the ability to auto-sync with Power BI visualizations. Not fancy, but still a big improvement for us. So far so good.

Looking around the Power Platform tooling, I could see some benefits with Dataverse Power Apps for managing data (then feeding to Power BI) based on what I've read. When we consider building this out more, we realize that some of our data is a bit multidimensional and/or has complex relationships with other data, and might be awkward to flatted into the constraints of a List. I imagine s set of Dataverse tables could handle that more effectively. I also know a db table would likely scale better as we add more data rows, though that really doesn't seem like a major constraint for our use case.

I've started experimenting with Dataverse for Teams implementing a few tables that we currently maintain on SP, but so far it seems... clunkier? Out of the box is seems like SP Lists can already do a lot of things for simple data entry and viewing that would take a decent amount of setup time if replicating in Dataverse PowerApps (implementing the different list views we use, creating a version control/approval workflow, even just making and entering multi-choice lookup columns, etc).

I guess I'm trying to understand the benefit of moving forward with Dataverse PowerApps to manage the data we track, versus sticking with SPO Lists. I feel like I must be missing something. Is it just a matter of scale - if I was dealing with "large" data, the choice would be more obvious?