Sentiment Analysis in Supply Chain Management

Sentiment Analysis in Supply Chain Management

Lincoln C. Wood (University of Otago, New Zealand), Torsten Reiners (Curtin University, Australia) and Hari S. Srivastava (PNG University of Technology, Papua New Guinea)
Copyright: © 2014 |Pages: 12
DOI: 10.4018/978-1-4666-5202-6.ch193
OnDemand PDF Download:
List Price: $37.50

Chapter Preview



The supply chain is a sequence of activities which are conducted in separate companies. Materials flow along the supply chain progressively becoming transformed into a product that a consumer wishes to buy. On the other hand, information is transferred from consumers to producers, impacting on the future delivery plans for the suppliers. Ideally, feedback about the behaviour of the end consumer should be driving the coordinated behaviour of the supply chain. When consumers begin to buy more products, the sensible initiative for the supply chain to do is to match the demand with an increase of the product output - in a process of demand and supply integration to improve operational efficiency (Esper et al., 2010).However, a well-studied phenomenon known as the bullwhip effect can often be observed. This is where significant fluctuations occur in the output of members along the supply chain, as members further away from the consumers tend to overreact to changes in the final marketplace; even in response to just small changes in consumer demand (Lee et al., 1997).This is largely due to the supply chain members’ lack of information about market-based activities. This phenomenon makes coordination and management of the supply chain challenging and creates additional costs and reduces chain responsiveness. Rather actual market-based information being shared along the supply chain, it is more common that suppliers take orders from consumers as an indication of market demand. As a result, many firms find themselves confronted with potential asymmetries of information along the supply chain and fail to respond efficiently on the market-based demand by responding to orders being placed by their direct consumers (O’Leary, 2011). This inability to determine consumer demand changes can be overcome using analysis of social media and opinions posted online.

Analysis of social media can be used to sufficiently predict social behaviour (Abbasi et al., 2012). The use of ‘sentiment analysis’ or ‘opinion mining’ can allow firms to derive an understanding of changes in consumer demand or preferences as expressed in social media, . Thus, ‘demand sensing’ of market-based demand allows firms to detect shifts or changes in trends in market-based demand (rather than orders from consumers) which then can feed into planning processes without requiring cooperation or coordination with other firms in the supply chain (Qin, 2011). Using business analytics approach of textual ‘sentiment analysis’ creates the opportunity to ultimately enable the prediction of sales. At present, commercial tools have been more readily applied to marketing and sentiment analysis rather than focus on supply management applications (Zitnik, 2012).

Sentiment analysis has great value for all suppliers independent of their distant from the actual end consumers. Suppliers closer to consumers, such as those in food and beverage supply chains, tend to have well-established information-sharing connections that help supply chain members to be more aware of consumer demand changes and respond accordingly; this is exemplified in the Beer Game (Sterman, 1989). However, all suppliers can benefit from both forecasting and sentiment analysis as a large proportion of their supply is directed through their supply chain; a change in demand for one product or category can have significant ramifications and thus should be monitored. In some cases, firms directly serves consumers (such as supermarkets in food and beverage supply chains) may be reluctant to share insights about consumer demand; in these cases, sentiment analysis can prove to be exceedingly valuable to suppliers in short supply chains.

Suppliers that are distant from the consumer are disadvantaged by the relative difficulty in retrieving or accessing consumer demand information; sentiment analysis can prove to be advantageous in this respect. This value is, however, balanced by the fact that their output through any one supply chain going into one product/category that is affected by consumer shifts is a smaller proportion of their overall output. Thus, a shift in consumer demand for a particular product/category may have a smaller impact on their overall business, yet sentiment analysis may be the only practical method for them to gain insight and forewarning that this shift is about to occur in the first place.

Key Terms in this Chapter

Text Analysis: The process of deriving meaningful information from the data and ideas expressed within the document. It includes meta-information, structural information, and content information.

Sentiment Analysis: Application of natural language processing, computational linguistics, and text analytics to identify and extract subjective information in text materials.

Bullwhip Effect: An observed trend of increasingly larger oscillations in inventory the further a firm is from consumers, in response to changes in consumer demand.

Supply Chain: Sequence of activities conducted in separate companies transporting a materials flow to consumers while progressively transforming it into the desired products. Information from the market flows to the producers.

Beer: A beverage that is fun to drink and enhances knowledge creation activities in universities around the globe.

Semantic Analysis: The elicitation of knowledge from documents, accounting for the context and understanding. The units that are extracted are arranged and grouped within meaningful categories.

Concept Retrieval: The ability to query a document and extract particular segments of text that match concepts or ideas provided by a user.

Concept: One or multiple words associated with a category that was generated by the abstraction of common characteristics from a range of particular ideas, while removing the uncommon characteristics. The remaining common characteristic is similar to all of the different individuals and represents the meanings of the ideas.

Sentiment Classification: Classifying the polarity of a given text in the document, sentence, or feature/aspect level; e.g., emotional states.

Complete Chapter List

Search this Book: