We talked about the common phrase Don't re-invent the wheel in the earlier posts (part-1 and part-2).
Re-inventing the wheel in tech is challenging, and you must be careful about when and how you do it.
Let's face it. You cannot avoid using external services while building your company, and have to use it at some point or the another.
Once you have reached a certain stage, it's a good approach to replace those services one-by-one by building them in-house to get cost savings and greater customisation and control.
Here are the 6 cases when you can, and should re-invent the wheel:
Team: Your team can handle the non-functional part like security, scalability, maintainability.
PMF: You have achieved PMF, and don't have any major updates in the short term, so you can focus on building services in-house.
Cost: The services are costing too much compared to building them in-house would reduce the costs significantly or the services just keep increasing their rates which is getting expensive for you.
Control: You need more features, customisation and control over the service, and the company cannot provide them (even at an Enterprise level).
Support: The support for the services is slow, or not good enough for your requirements, and it might create issues for your next phase of your product.
Compliance: There have been certain regulatory and compliance related changes, or now you want to be more compliant compared to earlier, and so it would be better to have it build in-house.
Like everything, it depends on what stage you are on and what your priorities are, so once must be careful while re-inventing the wheel.
I would like to conclude with a quote from Thomas Sowell,
There are no solutions. Only trade offs.