Graph databases excel that will benefit of treating relationships are first class citizens and handling complex data structures. Still, it is hard to ignore the maturity, reliability and talent availability for traditional RDBMS which make the decision harder to make. There is an inherent risk when embracing modern and newer solutions and so architects need to make sure they are making the right call.
When it comes to relational databases management systems, the techniques, tricks, and optimizations are very well documented and accessible for all members and abilities within a development team. Adding relationships to an existing relational database can be as simple as adding a join table or foreign key constraint. With the new join table or foreign key, relationships can be queried and are accessible. Essentially, getting the data into the system is well documented and relatively easy for the developer.
Data modeling techniques with graph technologies are more intuitive as they are closer to the human understanding of the data. With that translation being shorter, it should be possible to extract deeper and better insights from the information at hand. Once the technology had been mastered, it should be simpler to provide these capabilities to an application that will benefit from it. If there is a need to model and reason about relationships in your data, graph technologies provide a more seamless transition from human understanding to machine representation of your data and back.
The complexity of the data model becomes the next key driver. As the number of tables and relationships grows, the need for nested queries and joins follows and it becomes hard to ignore. Graph query languages are not match for SQL when it comes to talent availability and most likely will under perform in simpler application. But given their expressiveness and focus on addressing complex queries, overtime they become a real contender in terms of simplicity of the implementation and should go neck to neck when it comes to performance.
At the end of the day, graph databases are a risk that might be worth taking. Will the application or system benefit from its approach? If the answer is a crystal clear yes and the education, and talent acquisition challenges are manageable, then it is probably a wise decision. The good news that graphs databases will only get better as the space fills of innovative vendors which is much more exciting of what has happened in the last decade in the RDBMS world.