Key parameters are the baseline and the target scenario, requirements and constraints with regards to the evolutionary network roll-out, and assumptions on the comparability of the operation to other operations in the benchmark database, or any adjustments needed to make them comparable. Due to the relatively fast computation, a large solution space can be initially investigated. Inferior strategies can be excluded from further research, reducing the number of alternatives to be investigated in greater detail.

The bottom-up approach is a highly customized, specific business plan development based on detailed quantitative inputs for all genuine cost drivers. Instead of applying benchmark factors, each item, such as headcount, salary expenses, overhead costs, tools, office and site rental costs, etc., is computed individually and using a large number of discrete assumptions, e.g. on wage inflation, productivity, allocation of general and administrative costs, etc. Such a computation is resource-consuming and can reasonably only be performed for few, pre-selected scenarios, such as those resulting from a screening with a top-down model.

Top-down estimation model allows major conlusions

In our work, subsequently to identifying the most promising scenarios through the top-down method, we employed a bottom-up model to hand over to our client comprehensive, operator and location specific data suitable for executing the transformation. Here, however, we would like to focus on the top-down part, which already allowed us to draw the major conclusions, and which represented a novel approach in the way we implemented and applied it to our underlying problem.

Initially, the current situation is described and all potential target scenarios are defined. This is done by qualifying the service rendering for each combination of technical domain and process activity, as one of our five delivery options (or as “not applicable” if the process is not needed for a certain network part). Fig. 1 depicts an exemplary extract of the scenario table.

The various network operations processes have different weights within the operator’s cost structure, reflected in a breakdown of network OPEX shown in fig. 2 for our example case.