The reason is clear. All too often, companies take a wide-angled approach, losing their focus in the vast universe of available data, seeking to understand every possible piece of information out there. To take one example: in the effort to mitigate supply chain disruptions, companies go through vast amounts of data to quantify every risk that poses a threat, often without regard to how small the threat actually is. This approach, however, is a costly exercise that’s likely to yield lots of data but little meaningful understanding of where risks with the most impact exist.
A more fruitful strategy is to focus on the data that’s relevant instead of all the data that’s available. Determine the most critical questions you are trying to answer, and then work backwards to determine the information desired to make the best-informed decision. In the case of supply chain disruption, you might ask: Where do we need redundant suppliers or production capacity to minimize the financial impacts of a supply disruption? Should we stockpile any raw materials? Which ones and where?
It’s critical in this endeavour to abandon any preconceived ideas concerning the quality or type of data you’re looking for. The data doesn’t need to be perfect: Decision-quality analysis, not precise analysis, and multiple analytical perspectives are what’s called for.
Even with a focused approach, a lot of information and analysis will be necessary. Simulation modeling with “what-if” scenario analysis can help. A model consisting of four advanced analytical simulation techniques – discrete event, agent-based, system dynamic, and Monte Carlo – provides multiple perspectives simultaneously.
This unique multi-perspective approach addresses the challenges inherent in incomplete and uncertain information. In cases where a company wants to identify sources of potential supply chain disruption, the model focuses on only the nodes where a disruption would have a material impact. And for nodes where information is incomplete, the model treats them as “black boxes,” making assumptions about the frequency and severity of failures that can be revised as more information is obtained. The incremental and iterative nature of the modeling makes it easy to explore Big Data analytics concepts and develop proof-of-concept solutions.
This type of a focused, question-driven approach, coupled with multi-perspective modeling, allows companies to get a grip on – and value from – Big Data, not only to improve resilience to supply disruptions but to enhance event management, demand forecasting, and cost forecasting. How do you handle incomplete and uncertain information about your supply chain? What Big Data techniques have you used to address these issues? Companies must create dynamic simulation models to address their particular challenges.