Jewel Huang/ID 216216298
With the debating on big data heating up, marketers have complex feeling about social network data.
Waste or golden mine?
Social network data are readily accessible at low cost. Brands can establish direct interaction with a large group of customers and identify their attitudes on the company and specific products. The number of the interacting customers, as exampled in the “Facebook likes”, can largely exceed what interview, focus group or survey can handle.
Marketers can dig through rich information to identify how products interact with 5Cs: how customers expect, experience and evaluate products under real social contexts. Direct comparison between different products, both the marketers’ and the competitors’, can be established. The following table illustrates customers’ evaluation on different Starbucks in Melbourne CBD.
The social network data also reveal subtle customer behaviour. For instance Kraft once went through 1.5 million posts in social network and found 479,206 mentions of Vegemite with connections with avocado, tomatoes and roast meats. This led to Kraft’s campaign “How do you have your Vegemite?” that achieved 5% increase in sale (Fitzsimmons, October 10, 2013). Having Vegemite with avocado or tomatoes became a recommended recipe.
However a 2015 poll showed that marketers ranked big data, including social network, as the most challenging issues in market research. The issues include: How to dig through the big chunk of unstructured data; how to interpret date; how to find connections with segmentation groups; and how to quantify the qualitative attributes.
Fitzimmons suggests to start with questions (Fitzsimmons, October 10, 2013). As in any market research, the first step is to define market problems and to conceive hypotheses (Iacobucci, 2013), thereby reducing the noise and projecting the signal.
The next step is to draw meaningful information that answers the questions. This is called data mining. Although wild data can be intimidating, many technologies, e.g. text analysis and sentiment analysis, assist this process (Tsytsarau and Palpanas, 2012). Here is a mini-lecture on how it works.
In 2009 German Federal Election, Twitter messages showed, through sentiment analysis, accurate correlation between the times a party was mentioned and the election results (Tumasjan et al., 2011).
In another case, researchers analysed 7362 Twitter posts to find users’ attitudes to tobacco and electronic cigarettes. Correlation can be found with users’ personal experience and social image, but surprisingly not with their health (Myslin et al., 2013). Researchers used text analysis to estimate 100,000 Twitter users’ demographic attributes. The accuracy for gender, age and area is 84.5, 63.5% and 75.9% respectively (Ikeda et al., 2013).
Researcher can also draw quantitative information from customer reviews on the weights for product attributes and the evaluation score for each attribute – similar to perceptual mapping (Iacobucci, 2013)- and make correlation with product price (Archak et al., 2007).
While data mining has shown its potential in market research, this is not to say social media data would replace, but rather enhance, traditional marketing research methods. As an emerging area, the data analysis technologies have to further develop and the marketers need training to use the new tools. The ethical issues related to personal privacy also need to be regulated (Nunan and Di Domenico, 2013). It is expected that, with technological friendliness and effective regulating guidelines, social network data can turn into golden mine for marketers.
ARCHAK, N., GHOSE, A. & IPEIROTIS, P. G. 2007. Show me the Money! Deriving the Pricing Power of Product Features by Mining Consumer Reviews.
FITZSIMMONS, C. October 10, 2013. Big Data? Big Deal [Online]. BRW October 10, 2013: Fairfax Media Management Pty Limited. . [Accessed April 15, 2016].
IACOBUCCI, D. 2013. MM3 Student Edition, South-Western, USA, Cengage Learning.
IKEDA, K., HATTORI, G., ONO, C., ASOH, H. & HIGASHINO, T. 2013. Twitter User Profiling Based on Text and Community Mining for Market Analysis. Knowledge-Based Systems, 51, 35-47.
MYSLIN, M., ZHU, S.-H., CHAPMAN, W. & CONWAY, M. 2013. Using Twitter to Examine Smoking Behavior and Perceptions of Emerging Tobacco Products. Journal of Medical Internet Research, 15, e174-e174.
NUNAN, D. & DI DOMENICO, M. 2013. Market Research and the Ethics of Big Data. International Journal of Market Research, 55, 2-13.
TSYTSARAU, M. & PALPANAS, T. 2012. Survey on Mining Subjective Data on the Web. Data Mining and Knowledge Discovery, 24, 478-514.
TUMASJAN, A., SPRENGER, T. O., SANDNER, P. G. & WELPE, I. M. 2011. Election Forecasts With Twitter: How 140 Characters Reflect the Political Landscape. Social Science Computer Review, 29, 402-418.