Skip to content Skip to navigation

SMU Research Area of Excellence: Analytics for Business, Consumer and Social Insights

Introduction

­

Businesses have become increasingly reliant on analytics to make informed decisions. With the ubiquity of the Internet, mobile platforms and new social-sensing sources of digital information, there is immense potential to create unprecedented, deep and valuable insights for businesses and consumers.

Researchers from the Singapore Management University (SMU) are among the leading proponents and explorers on the mining and analysis of rich data resources from the online and mobile world. Their research activities involve the development of innovative tools that turn data deluge into meaningful knowledge, business and social intelligence.

From online transactions to social media to mobile data, the application of deep data analytics makes the research process iterative and dynamic. This enables the researchers to ask new questions, adjust their models and redesign their experiments to generate answers quicker than ever before.

The insights that they generate and their broad-based efforts in analytics-related outreach are starting to have a beneficial impact on the industry, government and society. Collectively, their leading-edge work in data analytics underpins SMU’s Area of Excellence (AoE) in Analytics for Business, Consumer and Social Insights (ABCSI).   

Analytics: Impacting businesses, consumers and social science

In 2012, IBM estimated that 2.5 quintillion (2.5 x 1018) bytes of data were being generated daily, and this figure continues to grow at an exponential rate. Our ability to produce data has far outstripped our cognitive abilities to make sense of it all. The challenge lies not only in dealing with large volumes of unstructured data, but also the rising velocity and variety of data. Be it the Web, the rapidly growing mobile sector or the Internet of Things, data are pouring in from multiple sources in myriad forms. People in many walks of life now wonder: How valuable is such data? And to what extent can we trust it to be a meaningful representation of the different contexts in which we live and work?

Researchers at SMU are interested in developing analytics tools to drive novel forms of consumer engagement and to enable the transformation of organisational strategies and business practices. To do this, they rely on emerging methods from interdisciplinary domains including Psychology, Sociology, Marketing, Logistics, Economics, Statistics and Computer Science.

They explore innovative ways of using experimental techniques to iteratively probe and refine the efficacy of resulting strategies in a highly interconnected and dynamically evolving world. This approach has deep and long-lasting implications for and across a variety of businesses (e.g., airlines, transportation, telecommunications, digital entertainment, pharmaceuticals, social, healthcare and municipal services).

Empowering consumer intelligence. By focusing on consumers, analytics support the development of highly personalised and context-aware services. The ability to analyse each consumer’s behaviour, as opposed to their aggregate behaviour via traditional forms of customer segmentation, supports the creation of products and services that are tailored to meet individual needs.

Creating value for business. Business leaders always want to know: How can I deliver the highest possible value in what my organisation does for its customers? Data analytics research serves to create data tools and analysis innovations that make it possible to extract unprecedented levels of value from the exabyte data pool. It also helps businesses identify what they do well, what can be improved and who will be inclined to pay for their innovations.

Data analytics can be used to predict changing markets, pre-empt operational problems and develop new insights for market-transforming product and service development. As such, businesses stand to gain a better understanding of their operations, consumers’ needs and the mechanisms for sustainable profitability.

Revolutionizing social science. The use of analytics to tap into huge social and human activity-related data sets has the potential to revolutionise social science research. Traditionally, the research cycle begins with the identification of a social problem and the development of research questions that can be answered by scientific methods. It is typical for researchers to identify an explanatory theory and frame a hypothesis for how things work. This is usually followed by data collection and analysis through a slow, expensive and non-scalable process. By leveraging the Internet, mobile devices and social data testbeds, researchers can now obtain real-time intelligence in an iterative and low-cost manner.

Interesting patterns of social behaviours that are observed through the lens of “steaming data” can help to answer the basic questions that are asked. It also has the capacity to allow researchers to dynamically reshape their research questions, as well as their empirical and experimental methods, to progress towards even deeper insights.

Leveraging deep data. While the idea of diving into big data may seem daunting to most people, SMU researchers view their mission in terms of discovery, insight and fun. The rich new sources in the realm of “data-at-scale” offer tremendous opportunities to know more about our world and all of its social activities.

The crux is to come up with user-friendly ways to collect and transform such data to analyse and present the information it offers, and then devise strategies and follow-up actions to create value from the insights it delivers. Clearly, the opportunity and significance of this work is to be able to harness the hidden value of the data for developing and implementing new business and social strategies.

These are the key themes that characterise the work of SMU’s data analytics faculty and researchers who come from different disciplinary backgrounds such as Finance, Marketing, Information Systems and Management, Sociology, etc.

Practical approaches to solving analytical puzzles

At SMU, researchers are actively developing practical business, consumer and social analytics through online experimentation, real-world experimentation and test-bedding at scale. Their approaches rely on the integration of advanced data mining, digital sensing, machine learning and statistical methods with social and behavioural science research (see Figure 1).

Through online experimentation, they can now perform interesting and realistic manipulations on variables to investigate causality in consumer behaviour and social interactions. This ability to conduct real-time experiments on actual companies and consumers was previously unimaginable. Today, it represents a mega-shift to the new paradigm of computational social science.

Figure 1. Analytics for Business, Consumer and Social Insights research at SMU (View larger image)

 

Professor Robert J. Kauffman from the School of Information Systems (SIS) analyses tens of millions of data points captured from the Internet, and identifies settings where real-world experimental treatments and controls can be set up. Through the use of advanced econometrics and biostatistics, randomised experiments and structural models, as well as counterfactual arguments, he has created a unique basis for studying a variety of interdisciplinary business problems.

He has applied these approaches to markets as diverse as transportation (train tickets), retail (supermarkets and cable TV content preferences) and e-commerce (group-buying and online auctions). He has also addressed questions such as: How effective are nine-ending prices (e.g., $99.99) and what behavioural foundations make the number ‘9’ an effective means for communicating discounts, quality and value? How do the amount, type and transparency of information made available to consumers affect their decision-making process? Will the analysis of deep data transform online selling with new forms of hyper-differentiated products and services? And will this lead to new approaches for marketing experience goods such as cable TV programming, video and digital entertainment services?

From his analysis of 1,800 drug brands from 30 major pharmaceutical manufacturers, Professor Srinivas K. Reddy from the Lee Kong Chian School of Business discovered that the recall of Bayer’s cholesterol-lowering drug, Baycol, adversely affected the company’s stock price and sales of its other drugs. To do this, he used state-of-the-art methods in statistical analysis to leverage the structure and temporal elements of the market announcement as a way to develop causal arguments for how brand failure may occur in unintended ways. Although this research was not conducted online, it nevertheless illustrates the breadth and inclusiveness of SMU’s research to embrace many ways of discovering new knowledge that shed light on some of the most important business problems of our time.

Another area of research that Professor Reddy has explored involves the inner workings of online art auctions. It has become possible to craft innovative and effective empirical research designs that reveal consumers’ social bidding behaviour in auctions for hedonic goods and services. With Internet auctions as his “research lab”, he is able to maximise the realism of his research designs, while making thoughtful choices about the precision of the new knowledge he can discover, and the generalisability of the findings he can produce – all in a context of truly rich and deep big data.

By combining the rigour of randomised controlled experimentation design with the ability to allocate online users to treatment and control groups in real-time, online experimentation studies can be more effective and affordable than conventional survey-based methods. Analysing the different behaviours of individuals in treatment and control groups helps researchers to attribute experimentation outcomes to the true causes.

To support large-scale online experimentation with minimal experimenter effort, SIS Professor Lim Ee Peng has embarked on the development Living AnalyticS ExpeRimentation (or LASER), a new platform for digital data experimentation that offers easy-to-use software components and an intuitive visual interface. LASER has already been deployed in several online experiments and was featured at the 2014 World Wide Web Conference in Seoul.

Professor Lim was awarded SMU’s Lee Kuan Yew Fellowship for Research Excellence in 2014 in recognition of his innovations in data analytics from the Computer Science perspective and his leadership of the Living Analytics Research Centre (LARC). Jointly established by SMU and Carnegie Mellon University in 2011, LARC is funded by the National Research Foundation (NRF) with a $26 million grant for a five-year period. Through partnerships with LARC, businesses have applied consumer analytics and participated in the test of the effectiveness of their marketing programmes.

In the same vein, test-bedding at scale allows researchers to ask questions and collect data in an unprecedented fashion. In particular, the ubiquity of mobile devices and their sophisticated in-built sensing capabilities present a unique opportunity to capture data on what individuals are doing. This presents opportunities for marketers who want to understand and influence consumer behaviour, logistics managers who want to streamline their operations, as well as commuter rail services providers who want to do social sensing related to congestion, crowding and service improvement.

While the aforementioned researchers tend to focus on consumer interactions observed on online platforms, SIS Associate Professor Archan Misra uses data from personal mobile devices to better understand the physical world behaviour and interests of consumers in public venues, such as shopping malls, airports and college campuses. He and his colleagues apply statistical stream-mining techniques over indoor movement traces and sensor data from mobile devices to determine whether a person is standing in a queue, or visiting a densely-crowded shopping mall alone versus in a group.

Such data help marketers understand how distinct groups in public spaces behave and what an individual customer is doing in near real-time, enabling retailers to improve customer experience by providing targeted, timely and relevant promotions and services. Businesses can use these insights to derive accurate estimates of queuing time, provide special rebates to individuals experiencing large queuing delays and create customised marketing promotions that are tailored according to the size of the groups.

Carried out under SMU’s LiveLabs research centre which is funded by the NRF for $9.9 million, research by Professor Misra and his colleagues has been tested in the SMU campus, Changi Airport, malls in Singapore, Korea and Japan, as well as food courts in India.

Apart from mobile phone sensors, social media represent another powerful resource for test-bedding at scale. Professor Lim’s research entails collaborations with businesses, government agencies and online service providers to uncover social and business insights from social media data. He and his research team develop novel data mining and machine learning algorithms that analyse millions of social media users to discover influential users and to predict churn of social media users.

They also conduct research that integrates multiple social networks with overlapping users, recommending items and social ties, modelling social and consumer behaviours, and discovering user-user relationships, while addressing other interesting applications. This stream of research has led to several important findings about users, their social ties and the communities in which they participate. Under his leadership, SMU has established its first real-time social media analytics platform capable of sensing and analysing terabytes of social media data.   
 

E-social science: wider trends in analytics

All three approaches – real-world big data analytics, online experimentation, and test-bedding at scale – are supported by cutting-edge tools developed by the researchers at SMU.

As communication networks involving devices, systems and services continue to grow, it is hard to imagine a future in which businesses and institutions do not expend major efforts to master analytics and data science. The growth of this discipline is changing the way we learn, sense and adapt to business, consumer and social contexts.

By initiating large-scale, real-world experiments on consumer behaviour across multiple digital platforms and via social media and mobile devices, researchers at SMU are paving the way for smarter decision-making based on sound empirical evidence at an unprecedented speed.

 

Back to Analytics for Business, Consumer and Social Insights

Last updated on 02 Jul 2015 .