Slavery was practised in many if not most human societies all over the world and continues to this day in many places, so it has taken a bewildering variety of different forms. This is really important to recognize if your only exposure to slavery is the Plantation Era slavery in the American South, for example. For example, slaves in pre-modern Korea (which I'm most familiar with) could form their own households and own property, even other slaves.

It can be surprisingly difficult to estimate slavery-related numbers in pre-modern societies. And these numbers would have changed significantly through time. In Korea, prisoners of war would have provided many slaves in earlier times, but this didn't last as wars became less frequent. Instead, in the late Middle Ages, the number of slaves increased for economic reasons, as plantations developed and people often sold themselves into slavery as agricultural labourers in order to escape economic hardship and taxation. This then became entrenched in a mostly hereditary system until government policies in the 16th and 17th centuries that sought to increase the tax base by limiting the number of privately held slaves led to a sharp decrease.

According to a census from the Daegu area, in 1690, 9.2% of the households belonged to the yangban aristocrats, 53.7% to commoners, and 37.1% to slaves; by 1858, 70.3% belonged to the yangban aristocrats, 28.2% to commoners, and just 1.5% to slaves. There are good reasons not to trust theses censuses (commoners would have been undercounted in the earlier period, and the yangban aristocrats were almost certainly massively overcounted in the later period), but they give you a good idea of the fluctuations through time.

This was because unlike in the American South, there was no ethnic or racial distinction between slaves and non-slaves. Just as one could easily be sold into slavery, one could also buy one's way out of slavery—or just run away and pretend to be a freeman. Slave registries could be destroyed in wartime.

In terms of numbers, at the height of slavery in the 17th century, perhaps up to 30-40% of the population were slaves, most of the rest being commoners with only a tiny yangban class. The yangban were the primary slave-owning class, but I don't know if all of them would have owned slaves. Many slaves were owned not by private individuals but by the government, working in government offices. On the other hand, some commoners and even some slaves would have owned slaves. Still, my guess is that at least half of the population were neither slaves nor slave-owners even then.