I have recently read the article «Deadline Driven Development» from Simón Muñoz. Quite a captivating one.
As in TDD, we create tests before developing the solution, he suggests setting a deadline before estimating, and then doing only what can be done until that deadline.
There are multiple perks to that approach: product managers avoid scope creep, and adding unnecessary features to the product being built. Other than that, developers will reduce the overengineering or creating too complex solutions, far beyond what is needed.
We will be reducing waste by building applications or services with just the two or three most important uses cases . Then, we can release our service, and see what was useful, what our clients require now, when they can use the product. From that point we can make more informed decisions on investing more, and better (or to cancel the product).

See the link to the original blog post (in Spanish): Deadline Driven Development by Simón Muñoz.