Should we adopt a component library?
Ask what it costs to leave, not what it costs to adopt. The adoption number is the small one.
Adopting is a sprint. Leaving is a year. That asymmetry should drive the decision far more than the feature comparison usually does.
- Build one real screen with it.Not the demo. A screen from your own backlog, with your own data.
- Price the exit.How many files import it directly? That number is your lock-in.
- Check the weight after tree-shaking.Measured in your build, not in their README.