There are tradeoffs between: performance, development speed, trainability (documentation), depth and breadth of developer community, long-term viability (foundation, backers), maintenance (upgrade path), API flexibility (decoupling, cohesion), standards compliance, vulnerability/risk (breadth), out of the box usability, accessibility, ...
So, for example, there are WYSIWYG tools which get like the first 70-80% of the serverside and clientside requirements; and then there's a learning curve (how to do the rest of the app as abstractly as the framework developers). ( If said WYSIWYG tools aren't "round-trip" capable, once you've customized any of the actual code, you have to copy paste (e.g. from a diff) in order to preserve your custom changes and keep using the GUI development tool. )
... Case in point: Django admin covers very many use cases (and is already testable, tested), but often we don't think to look at the source of the admin scaffolding app until we've written one-off models, views, and templates.
- Django Class-based Views abstract alot of the work into already-tested components.
- Django REST Framework has OpenAPI (swagger) and a number of 3rd party authentication and authorization integrations available.
- In a frontend framework (MVVM), ARIA (Accessibility standards), the REST adapter and error handling are important (in addition to the aforementioned criteria (long-term viability, upgrade path)) ... and then we want to do realtime updates (with something like COMET, WebSockets, WebRTC)
Similar features in any framework are important for minimizing re-work. "Are there already tests for a majority of these components?"