The scoring system has to reward this kind of behavior. But to be really good, it needs to do more. Jonathan Degann has written about the "big bomb" notion in scoring. A "big bomb" (put simply) is a high-scoring accomplishment that more than one player can strive for, but only one can achieve. Ideally the competition to claim the "bomb" involves some commitment from the competing players: the player who succeeds will see his investment pay off big, while the others will lose whatever they invested. This makes tension during the competition, and a big emotional pay-off when it's over: triumph or tragedy. Drama!
In the nameless rail game, I have been motivating players to build wide networks by designating six widely-separated junctions as cities (as opposed to mere towns). Players are rewarded with victory points for the number of cities in their network at the end of the game. Each additional city gives a bigger bonus than the previous. The first three or four cities are easy to get to, and the rewards are small. The fifth is harder, and gives a big slug of points. The sixth is hardest of all, and yields a really big slug of points. The board is laid out so that not every player can connect to every city, so the competition for fifth and especially sixth cities is fierce. This is one form of Degann's big bomb: for much of the game, players are racing to get to all six cities. Getting there first not only guarantees that you get the big reward, it also virtually guarantees that somebody else won't get there at all.
But I didn't want that to be the only way to get a big score, so I added another kind of bomb. This notion took me longer to come up with, and many iterations to refine. (I'm probably still not done with it.) It is the switchyard. Each of the six cities may have one switchyard built there. The player who builds it gets a big VP reward. This counts as a bomb because only one player can do it (per city), and because of the obstacles that must be overcome before the switchyard can be built.
First, you can't even start trying to build a switchyard until at least two players are connected to the target city. That guarantees competition. Second, you can't build a switchyard until you have made at least one delivery to that city of each of the four different kinds of good. Therein lies the race and much of the tension. Finally, after finishing the prerequisite deliveries, you need to come up with a big payment. Often two players will both finish the deliveries, and then it's a matter of who comes up with the cash first. The player who does will roughly double the value of his cash, while any who don't have wasted much of their effort.
I'd like to make that race more poignant if I can. There's more tension if the players have to invest something real in order to compete for the bomb. Currently, the only real investment is that players may forgo more lucrative deliveries in favor of ones that fill their switchyard prerequisites. And maybe that's enough. But I am considering making the players pay a piece of the switchyard cost up front, every time they "record" a delivery. The total cost of the switchyard is still the same, but you would have to pay some of it just to try to build the switchyard; and if you fail, you won't get any refunds.
What worries me about this notion is that it may become a disincentive: the potential loss may frighten players away from trying to compete for switchyards. If so I might try to lead them into the commitment gently: charge only a little for the first delivery, so it feels safe. But now the player is invested and wants to protect the investment, so if someone starts to compete he may be willing to pay a little more to keep up, and so on. I dunno, I'll have to try it and see how it goes!