Yeah. I could've worded it better. By "ejections" I meant how, when two planets/moon sized masses collide, rocks shoot out into space. But because black holes have so much gravitational pull, everything theoretically just falls in.
When two black holes collide, gravitational radiation shoots out into space. The origin of the radiation is in the dynamical spacetime outside each black hole's horizon, however. This is what the gravitational wave detectors operated by LIGO, Virgo, KAGRA, and others look for.
Similarly, the dynamical spacetime around a black hole not near any other black hole can couple with quantum fields -- even fields in a no-particle "vacuum" state as measured by an observer, for example one in orbit around the black hole -- with the result that Hawking radiation is produced.
Both gravitational radiation and Hawking radiation carry away energy (in the sense of ability to do work, per the "sticky bead" argument) from the environment immediately around a black hole. This in turn means that the horizon radius will be less than it could be.
So as a Hawking-radiating isolated black hole will tend to shrink (if it's not fed by hotter cosmic microwave background radiation, for example), the mass of a post-merger binary black hole will be less than the sum of the unmerged binary.
Just because things can't cross from the inside of a black hole horizon to the outside doesn't mean the horizon is always the same -- the horizon can grow and shrink dynamically when interacting with other self-gravitating bodies, with matter like dust or starlight, or with "the quantum vacuum".
