I like the idea of improving sustainability by using smaller Saturns. Even the "C2" and "C3", though, would have been large enough that Apollo would have been the only program using them, hence unit costs would be fairly high. So let's take the smaller-Saturns concept a step further and consider doing it with something in the Saturn IB class, about 40,000 lb to LEO. If the Centaur upper stage had been adapted to this vehicle for comsat launches (as proposed at one point) and if the the Air Force could have been talked into using it instead of the Titan III (not easy, I'm sure, but it would have been more rational than replacing the Titan with the Shuttle), it would have been possible to keep it in production economically.
A Saturn IB had the capability to launch a complete CSM/LM unfueled. Fueling the CSM/LM would in orbit then require about 70,000 lb of propellants; that's two Saturn IB launches. For TLI, the S-IVB stage will need to about 160,000 lb of propellants. That's maybe five more Saturn IB's, allowing for boil-off and transfer losses. So far we have a total of eight launches per lunar mission.
Of those eight launches, only one is "critical" in the sense mentioned in the Augustine report: that's the first launch, which carries the crew and high-value hardware. We need to be sure that that launch goes well. If any other launch fails, we just shrug our shoulders and launch another one. We will need one or two Saturns in reserve for each lunar attempt. Back-up vehicles not used can be saved for the next mission.
If we want to launch the crew on the first vehicle, we need some fuel in the SM so that the crew can abort if needed. Aborts after jettison of the LES involve the SPS engine, and the crew must have the capability to do a retro burn and return to earth if the mission is aborted before TLI. This probably means upgrading the Saturn IB a little bit. To avoid the risks of strap-on solids, the best way to do this might be to uprate the thrust of the H-1 engines to 250,000 lb (an growth option considered in the late 1950s), and stretch the first stage a bit. Maybe we stretch the second stage too.
If we can pull off one lunar mission a year, then on top of a half-dozen or so non-lunar launches (GEO comsats and Air Force payloads), we're averaging about 14 flights a year. At that rate, it might make sense to look into re-using the first stage.