That's a serious paradigm shift compared to what this (excellent) article describes.
This is a big part of their obsession with lowering costs to space. When the launch costs are not such a huge economic factor, you have much greater leverage with doing things like building, resupplying, or even engaging in interplanetary commerce.
Going to Mars may have been part of the story, but I doubt that it was a strong component of the decision making in the end.
Before that there would be a lot of work off Earth, including manned work, but it would be in space or on the Moon.
They key is Lagrange points. Each pair of bodies (Sun/Earth, Earth/Moon, Sun/Jupiter, etc) have 5 points where the gravitational forces from the two bodies balance out in a way that makes it possible for something to orbit that point, even though there is no massive body at that point.
Two of the Lagrange points are stable, meaning that if something in orbit around them is disturbed it still stays around that point. The other three are unstable. Disturbing something there will cause it to get farther and farther away.
You can use this to move things from Lagrange points of one pair (Sun/Earth for example) to Lagrange points of another pair (Sun/Jupiter say) very cheaply. Get it to the starting point, and then nudge it into an unstable orbit that will have it getting farther and farther away. We can calculate these unstable orbits well enough to pick one that at some point is nearly tangent to an orbit of the destination Lagrange point that moves toward that point rather than away. A little nudge them can move our ship into that latter orbit.
The catch is that this is slow. It might take decades or more to make the trip.
The way you would use this in a colonization program is to build a series of unmanned cargo ships. Say a new cargo ship is completed every year. It would be stuffed full of supplies the colony will need, send to an appropriate Sun/Earth Lagrange point, and nudged onto its journey.
Let's say these ships take 30 years to reach the destination. After we've been doing this for nearly 30 years then we'd send a ship with the colonists. That ship uses a fast but expensive orbit. It would only need to carry the colonists, the supplies they need during the trip, and fuel and supplies for an emergency return trip in case when they get there they find some reason that they cannot stay.
Note that it doesn't need to carry any material to actually build the colony, or food and water for the colony. All that is in the cargo ships that are now arriving yearly. (If we are sure that the cargo ships are making it the human transport ship could even omit food, water, and fuel for an emergency return. Those can be on the first cargo ship).
You'd want to build the cargo ships on the Moon, or build them in space using resources from the Moon, because getting from the Moon to Lagrange points takes a lot less energy than getting from the Moon to Lagrange points.
The most plausible path then is to greatly expand industrialization of the Moon and the space near Earth. Probably then expand that to include space bases at some of the Lagrange points.
Then it is time to start working on colonization.
Unfortunately we'd probably not do Mars this way. If I recall correctly the low energy Lagrange transfer orbits to Mars are particularly slow--over a thousand years if I remember correctly.