Initiative overview
The Water Sector Costing Forecast initiative for the Saudi Water Authority was created to provide a forward-looking view of plant costs by plant and by cost category. It was proposed by the Costing Department and delivered with support from the Data Management Office to address a practical challenge: historical cost data existed, but planning future costs remained largely reactive and difficult to scale consistently across plants. By using recent cost records to generate forecasts, the initiative helps finance, costing department, and operations teams make earlier and better-informed decisions.
The initiative seeks to improve budgeting quality, strengthen cost control, and make future spending patterns easier to understand. Instead of only reviewing past costs, the solution analyzes historical cost behavior and produces estimates for upcoming periods, allowing teams to anticipate increases in areas such as energy, maintenance, labor, chemicals, and other plant-level expense categories. This supports more proactive budget preparation, improves the accuracy of planning, and helps identify unusual cost deviations before they become larger operational or financial issues.
From an implementation perspective, the initiative starts by extracting costing data from the costing system, organizing and cleaning the data, and structuring it into consistent groups suitable for forecasting. The solution inventoried more than 136 cost items across multiple plants, removed older data that did not reflect recent cost patterns, treated missing and out-of-range values, and built a forecasting model using recent years of data. The outputs are then published to a costing dashboard so decision-makers can review expected future costs in a simple and accessible way.
The initiative has established a working cost-forecasting capability that supports forward planning for Saudi Water Authority plants. So far, cost data was inventoried across 17 plants and 136 cost items, presenting the result through a dashboard for decision-makers. Results were measured through data quality checks, test-set model evaluation using MASE, and forecast-versus-actual variance tracking. The main impact observed is better visibility of future spending by plant and category. Going forward, the initiative is expected to improve budget accuracy, strengthen cost control, enable earlier detection of cost deviations, and scale to wider plant coverage and more regular forecasting cycles.
