Schema for backing a multi-tenant SaaS

[Kenneth Cho]

I have an ArangoDB database backing an Apollo server GraphQL API that feeds data to a web application. Currently, the API and application use one database with one graph. This one graph has multiple vertex collections and multiple edge collections. 

We want to scale this application into a multi-tenant SaaS application where different groups of users use different sets of data with no overlap. I am considering changing the structure of the database to support this such that there is,

• One graph per organization of users
  
• One vertex collection per graph
• One edge collection per graph 
  

Does it make sense to structure things this way to support multiple groups of users, or do you recommend a different approach? (e.g., multiple databases, specifying ownership using a field in the data).