This paper provides an extended comparison between the use of customer class-based and area-based load profiling to assess the total revenues and the load deviations resulting from the consumption of the electricity users in a system partitioned into several territorial areas. Within each area, it is assumed that part of the users is subject to permanent metering, whereas for other partially-metered users only the energy values metered over a pre-defined time interval are known. The paper first illustrates various methods for allocating the consumption of the partially-metered users when the total load pattern is known for the whole system or inside each area. These methods, some of which are related to those used in national regulations of the electricity supply, result in different formulations of the allocation factors used to conventionally assign part of the total consumption to the partially-metered users. The effectiveness of these methods in assessing the total revenues and the total energy due to load deviations is then discussed with numerical examples. The total revenues are computed by considering a tariff structure with time-dependent energy prices. The load deviation assessment is carried out in a given time period, by comparing the load patterns of the partially-metered users, resulting from different formulations of the allocation factors, to the empirical load patterns simulated with a Monte Carlo approach. Numerical results showing the better effectiveness of the customer class-based load profiling over the area-based load profiling are presented.