Using a third-party design system is helpful, but there is a point when it outlives its use. What I’ve observed is that teams use third-party design system at different points in the product development lifecycle.

Teams may rely on third-party design systems when…

They are just getting started.
If you’re in a young company you’re focused on discovery, experimentation, rapid iterative prototypes and developing early workflows. Using an out of the box third-party design system just makes more sense when product details are ambiguous, you need to move fast and focused less on high polish.

As a stepping stone for establishing new libraries.
I like to consider libraries like Material as templates. They have some set rules, assets, etc. that can be augmented to create a more custom design system later on.

As support for an existing design system.
When the org has an established design system, but requires a specialized “sibling library” as a means to continue with project velocity when there aren’t resources, budget, or time to commit to building custom. For example, I’ve often run into core systems using a secondary library for complex data visualizations and interactive data tables that are reskinned to fit in with the primary system.

So when do teams diverge from third-party design systems?

When it's not enough.
If your old third-party system no longer can support customer needs it's time to reassess. At some point, the org needs custom components to support the use cases that the third party design system can’t provide.

When there's a need for nuance.
When there is nuance in how components are used or behave, growing complexity to reflect the needs of the teams that are using the system to design and build in specific domain. (This can eventually be reason to spin up additional custom sibling libraries to better serve specific user experiences, but that’s another topic).

People are the impetus for custom vs. out-of-the-box options

In an earlier post, I noted that design systems are products not UI kits. The above also illustrates the same. People — customers and feature teams — are what ultimately drives the need to adopt or diverge from third party design systems.