Comments on: Predictive Analytics & Probabilistic Planning

By: Klaus Spicher

Klaus Spicher — Tue, 19 Nov 2019 17:12:51 +0000

Hi Mr. Wilson, I would like adding another aspect. Identifying the drivers of business and utilizing drivers for forecasting represents a necessary but not sufficient condition. All factors which are not documented in the company’s files will not contribute to the forecast quality. The interdependence of identified and non-identified drivers remains in the dark of forecast deviations. In case the identified drivers explain 50% of the forecast, It’s a very acceptable result – depending on the business branch or products. Cleverly applied statistics enable utilizing inherent (unidentified) drivers representing a complement to Predictive Analytics – or the other way round. – I see the danger that the hype driver of Predictive Analytics only report the advantages but missing (potential) problems. One approach is not able managing the VUCA-Business World. The decision problem remains on the planners’ level.

By: Klaus Spicher

Klaus Spicher — Tue, 19 Nov 2019 16:48:57 +0000

Hi Mr. Wilson, I enjoy reading your papers. The logic of principles provide interesting common sense outcomes, as long as the the business world follows those principles. Nevertheless, for planners the transfer of the findings along “What-If’s” requires decisions influencing the the complete Supply Chain. Scenarios – a traditional approach for strategic planning, which no is broken down to the operational level due to Predictive Analytics tools, provide options for acting. But which option to follow? Predictive Analytics provides problems for the planner in operation. From all options finally decisions are required. Scenarios will support the In-House understanding the business. But, What-If, the planners next purchase order consists of several quantity options or a probabilistic range of quantities? Which company is interested discussing internal ‘scenarios’ with suppliers, also supplying competitors?