Yep, but if a "quickly" built Starship lander makes SpaceX a few billions, so they can invest it into the Mars program, then that may be a good option. For the same reason they are doing the Starlink network.
Related question: Just how many Starships do you think will be necessary to fulfill all our launch demands? Given that it's designed to be reusable and it has a huge payload capacity, perhaps there is less need to mass-produce them, since fewer vehicles will be able to do the same job as many expendable vehicles...?
Note that the killer feature of Starship, IMHO, is in-space refueling. This isn't a Earth-based heavy-life vehicle... it's a re-configurable spaceship able to explore the entire solar system through fuel depot staging (with Starships themselves serving as both tankers and fuel stations).
“I think there is a world market for maybe five computers.” -- Thomas Watson, chairman of IBM, 1943.
Then there is the idea of using Starships for very fast transport on earth. Fly from NYC to Sydney in half an hour. If that catches on, they may be replacing part of the airplane fleet.
Then there is lunar shuttle services, satellite transports and asteroid mining.
My guess is: a lot.
So it seems smart to say, "why don't they just build and test one, then figure out how to manufacture it at scale?" But to do that would be even harder because you'd have to redesign everything from scratch the second time around, while maintaining component compatibility... much smarter to design for manufacturing from the get-go and treat the test articles as expendable.
BTW, ULA is doing something similar with Vulcan & Centaur V. The design of the rocket itself is deeply related to the assembly line they are producing.