Statistics for effective and efficient processes
Forecasting is an essential part of business planning and involves a wide range of functional areas, such as marketing & sales, finance and logistics. A good forecast not only drives an efficient supply chain, it improves service levels and cash flow, and ultimately profitability. This is the fourth article in a series about sales and operations planning (S&OP) by dr. Freek Aertsen, academic director of the Executive Master of Operations and Supply Chain Excellence at TIAS.
Forecasts can be generated using statistics and/or judgement. A statistical forecast bases its projection of the future on results realized in the past by identifying trends, patterns and business drivers within the historic data. Judgmental forecasts, on the other hand, rely on intuitive judgements, opinions and probability estimates.
Decrease in costs
The use of a statistical baseline makes the forecasting process reliable, efficient, transparent, fast and objective. Depending on the possibility to centralize the planning process, a statistical forecast can be prepared very efficiently and eventually leads to a large decrease in planning-organization costs.
To generate a high-quality forecast, the demand signal for a specific product has to be differentiated according to the phase in the product life cycle (new, mature and end-of-life) and the distinction made between whether the sales demand was normal or part of a spike due to promotions, tenders and projects.
Statistics can be applied to support forecasting in many situations and offer the following benefits:
- Insights from the past
- Fast generation of different forecasts (see below)
- Scenario analysis and comparison
Three types of forecasting
Baseline forecasting for mature products is based on historical sales data and often uses trend and seasonal models. A high-quality statistical forecast allows companies to focus the enrichment process on those elements that really add value.
Promotion forecasting is based on historical sales and point- of-sales data, and promotion characteristics. A high-quality promotion forecast is generated (generally using regression models) for retailers and their suppliers to improve promotion effectiveness.
New-product forecasting is based on several internal and external data sources, historical introductions, volumes and characteristics, or social-media data. A high-quality new-product forecast can be used to improve the effectiveness of new-product introductions. Statistics (often multinomial logic) regression models can be used to forecast the full life cycle quantity, the initial launch quantity and the ramp-up profile.
Statistically generated forecasts very often show performance that can match or even outperform manually generated forecasts. If required for decision making, a statistical forecast is generated for all SKUs and markets. The aggregation level depends on the level of detail required for decision making / planning. The forecasted outcomes can be used to lend focus, e.g., specific knowledge on promotions and regional or local knowledge. The forecast is then enriched by adding specific knowledge of the local markets and customers.
Outsourcing has been on the agenda of nearly all supply-chain executives for the past decade. After all, third-party logistics and third-party manufacturing have enabled companies to focus attention on core competencies such as research, product design and marketing.
Outsourcing of the forecasting function is proving to be an increasingly popular option as companies continue to seek ways of improving their forecasting accuracy. There are multiple advantages to forecast outsourcing:
- Availability of specialized knowledge.
- Fast implementation – shortens time to value.
- Eliminates implementation risks.
- Economies of scale means lower costs.
- Continuous improvement due to investment in new technologies and skills.
- True, collaborative forecasting due to independent information broker.
- Best practice sharing.
The bottom line is that a specialist outsider delivers the best possible statistical forecast in terms of accuracy, efficiency and speed.
Read more: Step 1 to 3 in this series about sales and operations planning (S&OP) by dr. Freek Aertsen.
Optimizing your S&OP?
TIAS offers the Executive Master of Operations and Supply Chain Excellence which is a 11-module program, completed over a 22-month period. It offers you the latest theories, techniques and skills needed to successfully define and execute your company’s operations strategy.
or apply online.