Our theory of thermodynamics until now implements energy conservation but it does not predict the direction in which spontaneous processes take place in nature. Examples of spontaneous processes with a definite direction are
The reason that one direction occurs and not the
other is purely statistical. Take for example the adiabatic free expansion of a
gas into from a volume 
to a volume 
. The initial state of the gas
confined to remains a possible state of the gas at all times. However there
are far more ways in which I could distribute the gas throughout the available volume
than ways in which I can keep the gas confined to half of this volume.
The state of thermodynamic equilibrium to which we assume our system will
flow is therefore one which on average has the same number density of atoms
throughout the volume.
Similarly for the case of heat flow. We assume that the system will move towards a thermodynamic equilibrium and since the thermodynamic equilibrium will have a common temperature throughout, the hot region must get cooler and the cool region hotter and hence we have heat flow from the hot to the cold region.
In summary the assumption that thermodynamic systems develop thermodynamic equilibrium introduces a direction for spontaneous thermodynamic processes. Because we often consider systems which will not or are not intended to reach an ultimate thermodynamic equilibrium state it is useful to develop a tool which can predict the trend which exists as a consequence of the tendency to develop thermodynamic equilibrium.