Sensing demand signals in markets as complex systems: Wal-Mart case study.

Due to the turbulence of the markets and considering the markets as complex systems, the aim of this work is to focus on the relevance of managing the variables in forecasting, reducing the possibility to fail in market strategy definition; the high volatility in consumer markets for example, could justify the investments in conducting complex statistical frameworks. An empirical case study shows how could be possible to control many variables in order to optimize the prediction in demand management, in particular, in the retail industry. It is possible to obtain forecasts of demand using the history of sales yet, in the recent market turbulence could be useful to observe other factors. Forecasting demand is not just obtaining the history of sales, rather, observing other factors (external and internal) such as fuel price, consumer price index (CPI), temperature, markdowns etc., because, monitoring the variations of many more indicators, could increase the possibilities to better align tactics and future strategies that are of significance in the model. Analyzing weekly sales of 45 Walmart stores, additive non-linear model is introduced, that can examine the effect of each variable on the sales while holding of all other variables fixed. The Complex Adaptive System (CAS) framework is used to describe the complexity of the market and to justify the necessity to manage high number of variables. The complex networks and markets could be considered as CAS and are characterized by high number of variations in relations and interactions that affect the involved participants that are free to adapt their strategy to survive in a system characterized by the absence of a government body. The property of adaptability, comes from the capability to analyze the change in the external environment; the actors learn and survive organizing and re-organizing themselves in these kind of systems. This work presents a specific and practical approaches helpful to contribute to the adaptability of the actors to the complex systems, especially for the forecasts in consumer markets. The methods are alternative from the traditional approach and they are useful to make briefly structural analysis in demand forecasting based on inconsistent and volatile market. They are useful to predict the future including a lot of variables, at the same time, having a high flexibility and interpretability they contribute the adaptability of companies to the complex systems.