It is important to remember that the composition and penetration of various UE types will also have an impact on the final achievable cell throughput level. A high concentration of relatively low end UEs will result in low resources utilization efficiency, thereby bringing down the overall cell throughput.
This will certainly be depending on when the network is launched, pricing model of Operators as well as UE vendors as well as the form factors of UE offered at the time. The figure below summarizes the capability of UEs by their category. E.g. only Cat-5 UE can support 64QAM on uplink initially and will affect user uplink throughput.
Actual Cell Site Placement in Relation to Traffic
Similar to other cellular technologies, coverage provided by Macro vs Hotspot (micro/pico) cells and their vicinity to main user groups will have a significant impact on the final offered capacity. For example, limiting spectator traffic to a dedicated stadium only cell means most of its eNodeB power will be available to provide a better service rate instead of spending most of the energy travelling through air from external macro cell.
The actual capacity degradation due to the amount of users and their distance from cell antenna is highly variable and will also depend on the actual traffic distribution at the time. A general rule is the further the users are from the cell antennae, the lesser the amount of capacity a cell can offer. Capacity will be degraded even more if more users are to be located towards the edge of cell coverage and it is possible to have a capacity degradation of up to 25% in some situation.
User Traffic Mix and Call Modeling
LTE has 10 Quality of Service Classifications. The more freedom eNodeB has in user throughput allocation, the more likely the cell will have a higher aggregate throughput as the scheduler can adjust the resources allocation more appropriately based on radio condition. On the other hand, the more Guarantee Bit Rate users are present in the cell, the more likely the cell will have a reduction in its average aggregate throughput.