I think you can be fairly confident it will ultimately concern making money.
No doubt there will be numerous factors that have different degrees of influence over where will be preferable to deliver and in what order, so the 'quick hits' or more profitable locations are built earlier during the rollout.
Costs of deployment will be a major factor, but may not necessarily be a block if the percieved returns are sufficient.
Locations where there is competition from cable may well have some influence, but I doubt it would be a sole driver.
Number of premises passed will have some impact, so busy cabs that serve a high number of sites would be preferred over those that serve fewer.
In short, I don't think there's a simple "we will go here because of this reason", but rather a combination of different reasons will create the commercially sensible case for doing so.
Eats shoots and leaves.