Statistical Forecast is only a beginning
Your statistical forecast is only the starting point ..it is not the end-all and be-all.
So you are a professional forecaster. You have all the knowledge required to run a solid forecasting system loaded with features and functionality to produce best possible statistical forecasts! You are proud of yourself. Your R2 is in high 80s or even 90s so there is nothing more to it anymore. This is it! The ultimate forecast model that answers everything there is to be answered. So why is your forecast error creeping up? Why your MAPE is running in high 30s and Forecast Error on total business is equally dissatisfying? Why does Finance not look at your numbers and why the Supply Chain does not produce to what you are telling them via your perfect forecast? You have done everything possible in the world to produce this masterpiece.
The answer is in the fact that your perfect forecast is only one of many inputs into the final forecast the business needs to live by. The forecasting process is much more than just the statistical forecast. The statistical forecast needs to be adjusted for events that cannot be observed in past (R2 is a measure of fit of the statistical forecast to your historical data - in other words, if you rely solely on this measure, you are saying that the future will be merely an extrapolation of the past - nothing has changed in the market, no new competitor has emerged, no new private label product is "steeling" your market share).
Step one of the "Broad-View Demand Management model“ as developed and published by Oliver Wight International®, includes your statistical forecast ADJUSTED for new events that cannot be observed in past. The statistical forecast is still a backbone of the base forecast as it includes latest changes to the trend and seasonality, eliminating (with your help) the impact of past out-of-stock situations, out-of-plan promotions etc. However, it requires additional input from cross-functional players to accommodate the new activities that may be there to help close the GAP between the financial objectives of the company, and perhaps the reality shaping up after first few months of a new fiscal year.
The second step incorporates a very important part of any best in class demand planning process: communication of demand, collaboration with other key process stakeholders - sales, marketing, customers, finance, and supply chain. The key is to present your baseline statistical forecast adjusted for agreed to building blocks, activities during the last forecast/planning cycle as one of the key inputs into new planning cycle, (but not the only one.) All of this being relative to the last monthly cycle's agreed to consensus forecast.
Sales have their latest version of the market conditions, marketing has just collected the latest market share data and made adjustments to their media spending to correct the negative trend displayed in your baseline forecast. Key accounts you have been collaborating with via CPFR, have new messages in their promotional plans and in the way they see your products performing at the store level.
Step three involves the Influencing of Demand - in other words - once the team realizes there is a bigger GAP between the financial objectives of your company and the year to date sales, there is a need to move on with new initiatives, new promotions, out-of-plan launch of a product that may help to counter the current market situation, additional trade spend to be allocated to a particular, high margin, product group and/or customer - all of that to narrow and/or eliminate this GAP.
Step four involves addressing issues and their resolution when for example there is a sudden issue with the product availability. There may be a need to prioritize which customer gets how much of remaining stock, what will be communicated to the customers in terms of future product availability, possibly even the substitution for another, similar product.
So in short, the proper forecasting process leading to a Single Number forecast recognized throughout the organization uses your superior input as only one of the critical elements to be considered when working with other functional groups during a monthly planning cycle.
- 1. Planning Demand
- Involves more than just forecasting
- Communicating Demand
- Includes communicating the demand plan to supply and finance organizations and, increasingly, to supply chain partners
- Influencing Demand
- Includes marketing and selling tactics, product positioning, pricing, promotions, and other marketing and sales efforts
- Managing and Prioritizing Demand
- Includes managing customer orders to match available supply
Demand Planning process requires individual inputs critical to reach the consensus on what the final forecast should be for a particular monthly cycle. These inputs include Marketing, Sales, Business Plan, Statistical Base Forecast, Product / Brand Management, Supply Chain / Operations, ...
The Role of the Demand Planning Manager during the monthly planning cycle involves Business Plan, Economy, Statistical, Marketing, Sales, New Product, and Customer inputs to produce up-to-date solid Demand Plan the organization can work with, that satisfies your customers and suppliers and that provides viable solution to current market situation.
So, YES, the statistical forecast is a very important part of the company's planning process / inputs. However, it needs to be considered as only one of the key inputs into monthly replanning process, not the only one.