Other factors:
- is there tooling that will help me be productive with this language?
- is there a broad enough community of users such that finding helpful resources will be straightforward?
- how likely is it that the language I choose will be well-supported by its owners (or community) in 5-10 years?
- is the language expressive? can it be made to be efficient?
- does it appeal to me aesthetically?
- does it have good cross-platform support? (Windows, macOS, Linux)
- is it supported by build systems and/or CI products?
Through the filter of "stuff I know", my choice for most things at this point is F# on .NET Core/.NET 5. (and/or Fable).
For learning something new, I'd probably choose elixir with elm, or maybe Rust, depending on the application.
Should be web savvy (unless the front end will be native apps for iOS and Android, etc. then API savvy)
Scalable
Should support modern user management and encryption standards
There are a bunch of things you are going to have to list for your needs; What will your clients be like - will it be to companies with varying seats, or direct to users?; Does it need to have geocoding/geo-searching?; managing esoteric data?; Speed of response; capable of modular back-end stuff? Do you want to own it in-house or outsource development?; Compatibility cross-platform integration concerns?; Are there libraries already out there that you can utilize from the start?; Do you want to license code/data sources/API, or be independent, of external influences? How critical is the service, need to provide 24/7 emergency support?
This will likely bring you down to selections from the top 20 languages, then you need to look at how those work with your plans for development, like costs of licensing, availability cost of talent, and/or ease of programming. Then you have to get your fingers dirty or at least read up in the development processes and try them out and see if they are a fit or not.
If you aren't sure of much anything at all then figure out how to get from point a to b without expecting a finished product (create an internal proof of concept) so you can shop around for interest and see if you want to do such stuff.