Forecasters from different areas or departments have different business goals and therefore develop forecasts—such as the Strategic Business Forecast, the Tactical Sales Forecast and the Operational Supply Chain Forecast—according to their own aggregation levels, key figures, time horizons, and so on. For example, the marketing department focuses on brand or product group level and considers promotions in a medium term planning horizon, while the sales department forecasts for key customers and markets.
Consensus-based forecasting is a method that combines these different time horizons and forecasting levels from an organizational as well as from a tool perspective. Once all the parties involved in the demand planning process within a company have come up with their own individual forecasts, a consensus meeting is held. The result of this meeting is an overall demand plan that takes into account the business goals of all departments. The agreed demand plan is accepted by all parties. Moreover, all parties feel responsible for this demand and identify themselves with the output. Different demand streams get combined into "one figure".
Every party and department that is involved in the demand planning process has already created its own forecast; that is, sales people have defined a sales forecast based on key accounts, the marketing department has projected the impact of promotions, and so on.
The following process flow describes one example of how consensus-based forecasting can be handled in the APO system.
The following departments are involved in the demand planning process and deliver a forecast driven by the following factors:
The forecast is created for a combination of product and customer. It relates to the sales budget.
Naive methods and judgmental forecasting are used.
The sales forecast focuses on trade promotions, orders, syndicated POS data, competitive information, and customer information.
The goals of the sales department are tactical: to maximize sales and target actual customer demand. Thus, the sales forecast is profit-oriented.
The forecast is created for a combination of product/item and location on SKU level.
The forecast is history-based. Time series models and other models are used. Expert knowledge is also integrated.
The logistics forecast focuses on shipments, promotions, and material and capacity constraints imposed by suppliers and/or internal production conditions.
The goals of the logistics department are operational: to minimize production and sourcing costs, and at the same time fulfill production orders and sales orders.
The forecast is created for a combination of brand/product family and market zone.
Sophisticated models and causal analysis are used.
The marketing forecast focuses on events and promotion calendars, causal relationships, and syndicated POS data.
The goals of the marketing department are strategic: to increase demand volume, set demand streams, and bring stock or unused capacitiy to market through promotional activities.
The different parties involved in the demand planning process come together in a consensus meeting. The different forecasts are automatically aggregated and can be viewed on any level in the demand planning data mart. A specially designed consensus planning book is used for this purpose. The forecasts of the different departments are stored in separate key figures.
It contains the following time specifications:
Short to medium term
Once a week
It contains rows for the forecasts of each department. Manual changes can be made to these rows. The different parties discuss their point of view and, where necessary, make manual changes directly in the consensus planning book.
Another row in the planning book reflects the consensus forecast. It amalgamates the departmental forecasts. The numbers in this row are derived automatically using a special macro. This macro might calculate the average of the individual forecasts. Alternatively, it might weight the forecasts of each department differently.
The consensus forecast represents the agreed demand within a specific time horizon that will be released to and used by SNP for concurrent planning.
Where necessary, adjustments are made to the forecast models and/or to the macro.
The result of consensus-based forecasting is a one-number demand plan that has been agreed to and is owned by all stakeholders.