Because such a minimal change to the model ("there's just some more stuff we can't see") fixes a number of problems in one fell swoop (off the top of my head, g

Because such a minimal change to the model ("there's just some more stuff we can't see") fixes a number of problems in one fell swoop (off the top of my head, galactic rotation curves, observations of gravitational lensing, a cosmological model compatible with obervations of CMB fluctuations, structure formation, cosmic evolution). Theories of modified gravity can fix those individually, but as far as I'm aware, no alternative has been shown to be viable once taken in combination.