You're right, all of these processes require that you have all the bits in the right place at the right time and the cell does the right thing and no enzyme messes up too hard. So you shove as much stuff in as you can to maximize your chances When I would try to insert a mutation in a zebrafish, I would inject hundreds of fertilized eggs at the single cell stage, let them grow up, and then take a tail clipping to genotype. I'd be lucky if I got a couple successful mutations from those hundreds of eggs. It's definitely a numbers game.

Caveat: this was like 6 years ago, when it was relatively new. Success rates have likely gone up as the technique has been refined but general principle remains