Predictive analytics (PA) systems help in steering a company more proactively. However, managers often struggle in applying them. Focusing on a sociotechnical perspective from managers, we perform a morphological analysis to explore design dimensions of PA systems and their range of design options beyond “pure” IT approach. Based on findings from a literature review, we constitute essential design dimensions and complemented with results from cross-industry interviews we reveal designs options for each dimension. By means of a case study, we evaluate our artifacts’ utility and validity to finally discuss three lessons learned as follows: (1) Flexibility is key – PA systems are about creating new insights and becoming more flexible in turbulent times, in parallel relieve (local) controllers from their workload; (2) Simplicity beats complexity – start your project scope by applying time-series models. Then, if needed, upgrade toward more complex algorithms or even an ensemble of algorithms; (3) Corporate culture eats PA systems for breakfast – combine human experience and gut feeling with machine results, whilst internalizing the knowledge of consultants in your business department.