Sure, if you have already built the satellite, and it shares fuel between orbit raising and station keeping, it can be a big win. ULA convinced the US to do tha

Sure, if you have already built the satellite, and it shares fuel between orbit raising and station keeping, it can be a big win. ULA convinced the US to do that with GOES-16, recently. But if you're designing from scratch, building a bigger satellite and staging earlier gets more mass to GEO. That's called a "subsync" orbit and SpaceX has talked 2 customers into doing that with 3 satellites so far: Telstar 19V, Telstar 18V, and Hispasat 30W-6. https://www.reddit.com/r/spacex/wiki/launches/gto_performance https://www.reddit.com/r/spacex/wiki/launches/gto_performanc... In these latter 3 examples, the customer saved ~ $30mm on the launch by staying within the reusable F9 mass limit.