2,838 crowdsourcing and crowdfunding sites
Editor's Note: Crowdworx was kind enough to write up a case study for us looking at how one of its clients, Henkel, utilized social forecasting to improve predicting the demand for its wide array of products sold in Europe. Here are the results.
Henkel is a consumer goods manufacturer which operates in three different but related business areas: home care products, toiletries and adhesive technologies. Famous brands of Henkel are, amongst others - Persil, Pril and Schwarzkopf. The Henkel group has a workforce of approximately 48,000 employees in over 120 countries around the world, and is amongst the 500 most profitable companies.
As big as Henkel is, they were in need to increase the prediction accuracy in their S&OP process, and so they turned to Social Forecasting.
The supply chain management department at HQ was looking for a way to improve the accuracy of their sales forecasts of existing and upcoming products. The goal was not an incremental improvement, but a “step-change” in the forecasting accuracy.
The main reasons to change the existing forecasting model was the low forecasting accuracy and difficulties with evaluating the potential of new products. Additionally, price promotions performed by competitors influenced Henkel’s data-only based predictions and made them inaccurate.
In their search for better prediction methods to make that “step-change” in the supply chain Henkel turned to the international Social Forecasting vendor CrowdWorx. Henkel decided to try it out in one of their major markets – Eastern Europe.
Social Forecasting relies on the collective knowledge of a crowd – in this case Henkel’s employees. The employees have secure access to the Social Forecasting system via a web browser on the company’s intranet. Any participant is free to decide which products he or she wants to give predictions on. Once the participant has selected a certain product, the process follows this scheme:
The key to increasing the forecasting accuracy is the use of incentives in Social Forecasting. Each month the top 10 forecasters can win iPads and other valuable prizes. These top 10 forecasters also gain recognition. The difference vs. a survey is that participants are not rewarded for their mere participation but for their actual forecasting accuracy. These incentives greatly increased the forecasting accuracy as we will show below.
In the first few months Social Forecasting was running parallel with the previous forecasting process. The transition period provided valuable data that allowed a direct comparison of both Henkel’s method and Social Forecasting, under identical conditions. The chart below shows the results of this period. It clearly shows a significantly higher predictive power than the previous process. The average accuracy of Social Forecasting is 85.3%, while Henkel’s method achieved only 69.3%.
The data also shows that the accuracy of Social Forecasting is not only higher but the accuracy is also more stable compared to the previous predictions.
For Henkel’s supply chain management eight weeks of foresight are sufficient to adapt production plans according to the forecasted outcomes. This allows for timely responses, such as ordering a bigger amount of materials and packaging or the adjustment of machinery and production.
According to Henkel, increasing forecasting accuracy by only 15% would result in a positive EBIT effect of two or three-digit million dollar figures per year. From the perspective of the supply chain management the desired „step-change” in forecasting accuracy has been achieved and expectations exceeded. Henkel is now planning the introduction of CrowdWorx Social Forecasting in its four core markets – Germany, France, Spain and Italy.
- CrowdWorx is an employee-based crowdsourcing platform.