when does object storage become a good fit

Hey folks, thinking about data storage lately. We've got a lot of large files - images, videos, archives - that don't fit neatly into a traditional relational database or even a NoSQL document store. Object storage solutions like S3, GCS, or Azure Blob Storage seem like the obvious answer for this kind of unstructured data. But I'm curious about the tipping point. When does it actually make sense to introduce object storage into a system? Is it purely based on file size, or are there other factors like access patterns, metadata needs, or integration complexity that push you towards it? What are your experiences with managing and accessing data in object stores at scale?