Improving demand forecasting with the sales funnel and leading indicators

Thumbnail Image
Journal Title
Journal ISSN
Volume Title
School of Business | Master's thesis
MSc program in Information and Service Management
MSc program in Information and Service Management
Degree programme
In turbulent markets, demand forecasting is becoming increasingly difficult. Traditional quantitative models fail to anticipate market fluctuation, whereas managers' judgmental estimates tend to overshoot. Forecasting methods should be responsive to market developments to support proactive business planning. This thesis explores the potential of leading indicators and sales funnel in demand forecasting as a source of real-time market intelligence. Forecasting has been recognized as a key activity of customer relationship management (CRM), but little research has been done on how to actually use CRM systems for this purpose. The sales funnel is a valuable source of real-time market intelligence and prospective demand. Assuming the sales funnel of a firm follows the market conditions of the industry, forecasting visibility can be enhanced by using macroeconomic leading indicators of the sales funnel. How to embed such a forecasting model to the extant forecasting process in the firm is important. Literature from the fields of forecasting management, CRM and process improvement is used to build a framework for a case study. The case company is a manufacturer of lifting equipment operating in 48 countries, whose current forecasting accuracy is not optimal and is subject to judgmental bias from individual forecasters. A quantitative forecasting technique that incorporates the aforementioned sources of market intelligence is developed to improve its forecasting. The DMAIC model, famous from Six Sigma, is used to formulate a roadmap for improving the forecasting process on the whole. Unstructured interviews were conducted with the key forecasting stakeholders to observe current forecasting processes and accuracy, as well as the impact of potential forecast improvement on their operations. The sales funnel of the firm was analyzed against leading indicators in 14 countries or regions. If an indicator correlated strongly with the sales funnel of some country after synchronization, it was used for forecasting the funnel's values in-sample. The sales funnel, in turn, was converted to an order-level forecast with a simple optimization model. For most of the countries analyzed, a leading indicator was identified and applied successfully to forecasting the historical funnel base, at an average accuracy of 87%. Forecasting short-term demand from the sales funnel was 17 percentage points more accurate than any previous methods the case company used. The sales funnel was also used as a reality-check for further forecasts to identify major discrepancies, resulting in an improvement of 21 percentage points in historical accuracy. Experts at the case company estimate such an improvement to save at least 10MEUR in costs annually, primarily in capacity planning and procurement. The sales funnel-based forecasting model is more market-responsive and brings strategic value to the CRM system of the company. Applying the adapted DMAIC model to investigate the forecasting process at the case company revealed further strategic areas of improvement, giving the firm an action plan to improve its forecasting on the whole, taking a major step toward more proactive planning.
leading indicators, demand forecasting, market intelligence, sales funnel, forecasting process
Other note