To market better, know your customer better. It is a no-brainer that successful enterprises must anticipate changing marketing conditions, be responsive to customer needs, and compete by offering superior value. Especially when there?s a hybrid online audience nowadays?a shopper, a TV watcher, a tweeter and a texter?accessing the internet through mobile devices, tablets, PCs and access points outside of home and work locations. Understanding consumer behaviour so that the end-product or service may be even better than expected, is a prerequisite for business growth in the long term.
How can enterprises improve their understanding of consumer behaviour? Can new, evolving technologies improve firms? strategies and decisions? ?Definitely yes,? says Yahoo! Research senior research scientist Sharad Goel, who works in the microeconomics and social systems group in New York. ?By means of an emerging science called computational social science (CSS), we can show what consumers are searching for online and hence predict their collective future
behaviour days or even weeks in advance.? On a recent visit to New Delhi, Sharad shared some insights on this new form of science, including the new trend on the viral diffusion of information.
While we?ll demystify CSS later, it is imperative to have a holistic view on how enterprises can predict consumer behaviour with Web search. As people increasingly turn to the internet for news, information, and research purposes, it is tempting to view online activity at any moment in time as a snapshot of the collective consciousness, reflecting the instantaneous interests, concerns, and intentions of the global population.
Understanding and quantifying human behavioural pattern has
become far easier today than what was possible just a decade ago. Most of the apparel brand, travel agencies etc observe and reach customers by studying their behavioural patterns on buying, shopping, location preferences etc.
Building a successful business clearly requires lots of intellect, time and money, but other elements play a crucial role as well. For instance, predicting consumer behaviour with Web search?a passion for Sharad nowadays. And he knows this pretty well as his research interest revolves around empirical and theoretical problems at the intersection of computer science and the social sciences, particularly questions motivated by sociology and economics.
With the increasing availability of network and behavioural data?from what we buy, to where we travel, to whom we know?scientists are now able to observe and quantify
social processes to a degree that would have seemed impossible just a decade ago. These new microscopes into human activity not only have implications for the social sciences, including economics, sociology, and psychology, but also raise challenging computational questions in large-scale data analysis.
?Recent work has demonstrated that Web search volume can ?predict the present,? meaning that it can be used to accurately track outcomes such as unemployment levels, auto and home sales, and disease prevalence in near real time,? Sharad says. Quiz him further and he elaborates: ?Specifically we use search query volume to forecast the opening weekend box-office revenue for feature films, first-month sales of video games, and the rank of songs on the Billboard Hot 100 chart, finding in all cases that search counts are highly predictive of future outcomes.?
According to Sharad, the new sources of data available on social networks, news sites, Web search and sensor networks can be used to predict consumer behaviour. CSS is a computer science that is concerned with answering questions of social science but from the perspective of large data and computational tools. It is about learning how people behave, particularly how groups behave, using ?big data? and large scale observational data.
For example, how does an idea of product spread through individuals and reach entire population? It has been very hard to answer this type of question through traditional social science because the latter uses surveys. Surveys are often good at getting in people?s opinion, but are not good at predicting people behaviour. Laboratory experiment is another common approach used for answering social science questions. But this process doesn?t have kind of
external validity. On the other hand computational social science seeks to address these types of questions by using large scale observational data and tools from computer science and statistics.
Yahoo! Research team works intensely on the CSS techniques to understand the dynamics of its users? search patterns. This
research helps Yahoo! to understand its users? browsing patterns, and provide a more personalised experience in areas such as search, advertising, videos and homepage experiences.
Currently, there are only handful of researchers who are working in this area. In five years, this is going to be mainstream discipline. It is going to change the way social scientists work. It would be a significant advantage to be familiar with computing and statistics for those working on big data in social science disciplines. ?Many companies have already started advertising for data scientists. You know how do we optimise our marketing campaigns but I think that is going to shift towards longer term directions which is what Yahoo is doing, this is what our mission is to do long term research and using computational social science,? the Yahoo! researcher says.
From enterprises? point of view, they can throw out their old method of merchandising by assumption. Leveraging the power of the internet to better target the consumers with the right products is the way to go.