Influential Non-technical Factors for Shared Costs

by Neil May, Wish Star

Vendor pricing strategies – complex, ever evolving software and service licensing models from multiple vendors to flow through predictions and usage accounting.

Centralized Procurement In Multi-National Companies – due to corporate strategy for controls and/or seeking economies of scale from full company buying power

• Pulls all costs into one regional billing point, persistent monthly stream of usage data to couple with billing data for real-time analysis, validation against exceptions rules such as tag validation, clean digital identity integrity with budgets/cost centres and bill to entities in group, contracted rates within contracted commits and/or accurate overage rates etc, etc, etc…
• Validated charges to be re-rated (+ any internal finance admin charge) and cross border currency forex applied from what base, local tax jurisdiction treatments service by service (i.e. SaaS vs. IaaS vs Telecoms vs Network and accounting for withholding taxes incurred where international tax-treaties don’t support (i.e. WHT from Brazil to the UK is ~30% – who takes this cost/cashflow hit? how to account/apportion? How does this change corporate policy decisions to decentralized billing (from local vendors) to centralized vs. reduced economies of scale?

Exceptions – E.g. Pandemic causing sharp fluctuations in resourcing (human or otherwise with Furlough, contracting, burstable cloud services) creating exceptional usage patterns and the need to ensure flexible subscription plans are scrutinised for integrity with joiners/movers/leavers and project/process suspensions etc.

Pattern spotting – Building up the significant data lakes with triangulated data points between related services, allows for build up of benchmarking industry costs and utilisation models for total cost analysis of LOB App and solutions for apportionment. Data science and analysis can surface insights to help drive down costs as well as apportionment norms and the shared cost data can also aid prescription for quality of experience for user adoption etc etc. (different network provider data, compute data etc – optimising cost to performance for better apportionment)