Yup. That's a pretty freakin' brilliant concept too. Theoretically you launch less and less important payloads (like water or raw materials) until you finally fail a launcher. (Obviously going THAT far is not okay, but you get the point) That concept goes hand in hand with a current push in the industry to develop [automated] manufacturing in space: A significant amount of the work that goes into design/build/test of a spacecraft is launch environment survival...and so a significant amount of the cost goes toward that necessity. If you could, say, print your structure or circuit cards or wire harnesses in space, you just need to launch dumb raw material for those components. You also no longer need to design those components to survive a launch environment, so you can better optimize them for the mission at hand. You're also much less concerned about mass/volume, so you can deprioritize those aspects of a design trade. Its all good for reliability AND cost.
You could also aggregate a number of low cost spacecraft in a 'less reliable' launch as well, and that can really be a massive catalyst in lowering the cost of space. The biggest downward spiral of the space industry is the $$$ involved. More $$$ means more risk which means more mitigation/insurance, which means more $$$, and so on... Rockets are cool and all, but they're just an expensive way to travel. The real mission is the thing that's getting put into space, and right now the price of the rockets are still high enough that they're indirectly driving the price of the mission hardware. Elon is doing well to bust out of that circular reference and step function the cost of getting to space; some of us are doing our part to step function the cost of the space missions themselves.
The hard part about the later is most manufacturers are public companies where profit margin is a legitimate requirement. Elon is smart enough to keep profit motives out of his shop, and lucky enough (because he's smart enough) to have deep coffers backing him up. In the end though, its going to be an upward spiral where lower cost on either side drives the total cost down. Marginally Related Soapbox:
There's a legacy concept from Gemini that has perpetuated itself through the space industry at large, where failure is not an option. While understandable for manned missions, that's the reason why you hear of billion dollar DOD/JPL spacecraft that take a decade to build and eventually launch outdated technology. One can quite literally build the exact same spacecraft with the exact same performance, in half the time for half the price, by shaving just a bit out of the reliability budget. One can then build a second spacecraft with the other half of the money, with better performance (due to technology progress), and launch it when the original spacecraft would have launched. One then ends up with two spacecraft on orbit, with collectively more than twice the performance of the baseline, AND collectively similar if not better reliability ('strength in numbers', so to speak). Not to mention that the more you design/build/test the more you hone your concepts/processes/etc., so the more reliable your product becomes.
The commercial market is coming around to the idea of reduced part/component reliability driving reduced costs, but there's still so much money involved (~$200-300M+ for a big GEOcomm and launcher) that there's plenty of hesitation when push comes to shove. Heritage is still a really strong concept. The government (DOD and NASA) are barely opening their eyes to the idea of lower cost for lower reliability, but nobody of any importance is ready to risk taxpayer money ( = their job) on a program that doesn't fully CYA themselves with a billion dotted I's and a billion crossed T's. Its going to be on the commercial industry (propped up significantly by a low cost SpaceX) to basically shame the gub'ment into getting on board.
