A a reference point from the 2016 IAC, the first thing to note about Mars Colonial Transport/Interplanetary Transport System (MCT/ITS) is that it's intended to carry 100 people to Mars, plus "luggage," everything a colonist can take in his or her mass allotment.
The reusable MCT is capable of putting 300 tons in LEO. Since it's refueled in orbit, this means the plan is also for it to put around 300 tons on Mars. The exact mass depends on the launch opportunity, and ranges from 200 to 450 tons, depending on the delta-V required (200 tons with a 6 km/s delta-V, 450 tons when 4 km/s works--which can be a favorable planetary alignment, or when a slower trajectory is acceptable). There is also mention that in "expendable" mode, it can put 550 tons in LEO. Presumably, that equates to putting up a large space station module or something else that doesn't need to land, ever. (An Aldrin cycler comes to mind, but that's a topic for another day.) An expendable version would probably just use the second stage engines and tanks, with all the personnel and cargo space replaced by the payload. (We see the same thing with Falcon 9--the payload can be a reusable Dragon, or it can be satellites. The cargo a Dragon delivers is much less than the size of a satellite a Falcon 9 can put up.)
In his IAC presentation, Musk says his goal is to put one million people on Mars within about one hundred years. Throughout the presentation, the discussion mostly revolves around putting large groups of people on Mars at every launch opportunity--at least the 100 passengers on one MCT, but preferably several MCT's leaving together. So, that also makes clear he's talking about how large numbers of people go once there's an infrastructure in place to receive them. In the press Q&A afterwards, and other venues, he's mentioned the initial base ("Mars Base Alpha") might start with around 12 people. (The Q&A was edited out of the available online presentation, presumably because the questions were absolutely horrible. Sorry if one of those questioners is reading this now. The transcript is available at http://toaster.cc/2016/10/04/IAC_Press-Conf-Transcript/.)
Our focus here is on that much earlier stage, which Musk also clarifies is not completely designed. To make our design goal clear and keep it laser-focused, we spell it out in the site's tag line. I'm a fan of mission statements, so:
Mission: To establish a permanent, self-sustaining colony of 100 people on Mars, with the capability for further expansion.
The number is entirely arbitrary, of course, but there's something deep in the human psyche about round numbers, particularly when they jump to a new level. A question to ponder: would you rather be the 100th person on Mars, or the 99th?
To set the stage, it's worth spending a couple paragraphs establishing some reasonable assumptions about what SpaceX had planned for early MCT development and testing. That will hopefully avoid cul-de-sacs later. The presentation showed several years of testing, so we have a fair amount of latitude in deciding what technologies we can safely assume have been proven reliable.
Before the first MCT heads to Mars with any kind of load, several precursor steps will have naturally occurred. 1) Unmanned launch to orbit, followed by recovery on Earth. 2) Unmanned launch of MCT and tanker to orbit, with on-orbit refueling. That raises the question, "What to do with all that fuel?" which is a problem no spacecraft since Apollo 9 has had to face. 3) To use up that fuel, how about a trip around the Moon, followed by an accelerated return to Earth, with reentry at Mars-return speeds. Wait...there's a plan for that.
Musk hints that there may be some "local" opportunities with MCT; he mentions rapid intercontinental flight specifically. (Essentially, the spaceship alone becomes a realized version of the X-30 National Aerospace Plane, or Orient Express.) But, there's another economic opportunity that we'll need to exercise anyway to prove out the life support: multi-week trips for small groups of people. Sorry, Jeff Bezos. You want $250,000 for a 20 minute suborbital flight? For the same price, how about a 2-week trip orbiting the Moon?
Long before 100 people ride on a ship, we'll want to take 10 or 20 up to ensure everything works as expected. (For consistency, let's use Musk's suggestion of 12 people initially, and make it 12 to 24 during development and near-Earth test.) To be blunt, the only way to be sure your zero-gee toilet works right is to put it in a zero-gee environment and use it for a while. Lunar orbital trips are one possibility; as we extend the test missions, with a large amount of fuel available, missions to a near-Earth asteroid would be worth examining.
Both of these mission concepts were put to the SpaceX CEO, Gwynne Shotwell, in 2014 on the Space Show (http://thespaceshow.com). At that time, she indicated SpaceX had little interest in lunar opportunities, and was unfamiliar with NASA's plans for visiting near-Earth asteroids. (In fairness, NASA's plans aren't exactly crystal clear.) However, I submit these aren't distractions; they're an opportunity to prove out technology, make a little money on the side, and generate tremendous buzz.
It's enough to make one wonder--on the day SpaceX starts really bending metal on the MCT or something similar, what happens to NASA's SLS? Blue Origin's New Glenn is at least partly reusable, but SLS is expendable and costs $500 million per vehicle.
At the very least, perhaps SLS needs a LOX/methane upper stage, refillable in orbit?
A thought for another day.