Short answer: it depended on time period and location. It was ad hoc in every case. In general, the Romans set aside land as ager publicus (land belonging to the Roman state) and left the rest to the locals. How much depended on many factors and was different for every place. Sometimes after defeating an opponent they didn't do this at all, but instead set up a "client king" who was friendly to their interests. If they did set aside ager publicus, they might leave it alone, allow the locals to farm it, or put a colony of some kind there. Often, if there was already an urban center in the area, they would put a colony on top of it (like at Colchester, in England, which had been an Iron Age British oppidum).

Long answer: the Romans did not begin with, and never really developed, a systematic method for controlling new territory. The word provincia in Latin originally meant simply a military obligation; later it came to be associated with a military command in a certain place; and then finally with a geographically-bounded area under a Roman governor's control. The conception of territory for the Romans was therefore evolving as they acquired it. The origins of Roman control of conquered territory goes back to 338 BCE, when Rome settled things with the Latins and really began to be aggressive in Italy. It was after 338 that the system of ager publicus began to be extended.

Two different major means of dealing with ager publicus emerged. First was the "Latin Colony." These were established beyond Roman territory, frequently in areas where there were still hostilities. These foundations had to be defensible and usually included a stout circuit wall. Livy reports that around 2,500 colonists were sent to Cales and Luceria (Livy 8.16; 9.26) and 6,000 to Alba Fucens and Cremona (Livy 10.1). These are figures for adult males, either Roman citizens or not. It isn't clear what happened to the former owners of the lands these colonists would occupy, but we can assume nothing good. The Latin colony, once founded, was intended to be self-sufficient and self-governing (some even produced their own coinage). They were to provide a stronghold and source of manpower. The colony was, of course, under the Roman thumb so to speak, received de facto rights as a "Latin" charter, and was culturally Roman through and through (even if mainly non-citizens were sent there to found it). As the colony grew, it added public buildings in a typical Roman way: a forum, temples, house for the local Senate (curia), a place for the popular assembly to meet (comitium), a basilica for public business, and retail installations. If the colony was installed atop a pre-existing city (like at Paestum in southern Italy), these elements were "forced" onto the urban landscape as much as possible.

Second was the "citizen" Roman colony, also called a maritime (because many were founded on the coast) colony or simply "Roman" colony. These were distinct from Latin colonies in that their settlers either started with, or took on, full Roman citizenship. These were not set up as autonomous new city-states like Latin colonies, but were rather extensions of Rome itself. This type, OP, is approaching what you might imagine as direct Roman control of new territory. These places were more like garrisons than new cities, and were much smaller -- usually 300 families (Livy 8.21; 34.45). An example would be Ostia (the future Roman port), and its initial size as detected archaeologically seems to support Livy's small number. In general, these Roman colonies seem to have been placed in strategic areas -- on the coast, perhaps, to guard against seafaring threats (Rome was very slow to develop her own naval forces). Minturnae was placed on the Appian Way right where it crossed the Liris river on the way to Capua, for example.

I think it is best to think of Latin and Roman colonies as the ancient equivalent of medieval castles: they were strong points, garrisons. Rome would put them down, then connect them to roads. It was by these types of colonies that Rome wore down the Samnites, their bitter rival in domination of Italia. Cicero remembers these colonies of the middle Republic thus: "...it is worth remembering the careful practice of our ancestors, who so established colonies as a defense against danger that the colonies seemed not so much to be towns of Italy, but rather bulwarks of empire." (Cic. Agr. 2.73).

There was a third method, often called viritane settlements. These weren't official colonies; instead, the ager publicus was just assigned piecemeal to individuals or families. I have to admit that I know very little about this type or how it worked -- it is definitely a gap in my knowledge. I'm not sure how they defended themselves (I imagine early American homesteaders, far out into the western wilderness and surrounded by hostiles). I know that the major role of viritane settlement was to bolster manpower in particular regions, so that if an army needed to be called up in a hurry, there were strapping young Romanized lads available.

This system lasted until the early 2nd century BCE and was used liberally. But, at that point problems emerged. Latin colonies start to have fewer colonists, and some Roman colonies don't last (they are abandoned)(Livy 34.45, for example). I don't know the reasons why, but the Romans tried to fix the problem by combining the two types into something new: a foundation with the manpower of a Latin colony, but with the colonists having or receiving full Roman citizenship. Mutina is an example of the new type, founded 183 BCE with 2,000 colonists (Livy 39.55). Eventually the Romans stopped making the Latin type altogether after about 180 BCE, for reasons unknown (Livy doens't help here). There were many Latin colonies already in existence, however, and they continued and flourished. By the middle of the century, some had become very powerful cities in their own right. This led to clashes, most famously at Fregellae in 125 BCE. Most of the rest remained loyal, however. When the non-Roman Italians went to war with Rome in 91 BCE (the "Social War"), the Latin colonies remained loyal to Rome (and probably were largely responsible for Rome winning that war, for the record). After that, they received full Roman citizenship (but small consolation, because everyone else in Italy did too). There were very, very few colonial foundations of this old type after the Social Wars. More on that below.

So much for territory in Italy. Things become a bit more complicated when Rome ventures outside of Italy, and that happens first after the conclusion of the First Punic War, when Rome suddenly came into possession of Sicily (and Sardinia). We don't know exactly how Rome organized things in Sicily immediately after the war was over. It is likely that, at first, the Roman Senate just did things community by community via messengers. Perhaps a Senatorial envoy visited the island on occasion on this or that task. At this point there were still only four senior magistrates at Rome: two consuls and two praetors, and they were busy doing not-Sicily things. This was different from the war years, when a magistrate of some sort had been present or nearby to direct things in Sicily. With no authority there, problems arose: disputes about jurisdiction; disputes about tribute collection and assessment; and general disorder, one of Rome's greatest pet peeves. In the past, Rome had managed the various populations of Italy with ad hoc treaties. But for Sicily, a simpler solution was created: send back a figure of authority to the island to act as a surrogate war-time commander.

Thus the number of praetors was doubled from two to four, and two of them were to now go to Sicily and Sardinia as their provincia. This was the birth of the system of Roman "provinces." Territory in overseas provinciae was, as it had been in Italy, mostly left to the locals, with some set aside as ager publicus. I have done a good bit of hunting and can't find any specific numbers for how much was set aside. A lot of it was driven by the founding of coloniae in these places, especially by Caesar and then Octavian, who settled hundreds of thousands of soldiers in provincial colonial foundations. These colonies were a little different than the Latin and Roman colonies I talked about above, at least on paper. They were almost entirely discharged soldiers, were always Roman citizen foundations, and could be quite large (though I don't have specific numbers). In this case, the colony could be new or could be installed on top of an existing town (like Colchester, Corinth, or Carthage). Their basic function was still the same, though: provide a bastion of Rome in an area which might still be a little hostile, and provide manpower should things go to shit. The colony came with an attached agricultural area, called the chora in Greek, and this area would be meticulously divided up and distributed. This is called cadastration, and we have a few Roman representations of this kid of thing (the famous Cadastre d'Orange, for example, though this is from the 2nd c. CE and not from a colonial foundation).