The dividends of improving best practices for social media research
Jacqueline Anderson, Gina Pingitore and Miriam Eckert
J.D. Power and Associates
Social media is continuing to gain traction in the world of market research. While this is an exciting development that offers new opportunities for data collection, there is a potential downside. With so many tools available to so many different types of users, the "wild west" approach to social media data access may be leading us to questionable results. The ability to trust social media data that is delivered by market researchers will be called into question once stakeholders become aware of this issue. Without methodological rigor, results from social media research can at best simply be unreliable, and at worst lead decision-makers to draw wrong conclusions.
In this paper, we will show that without well-established and proven guidelines on query construction and data extraction, very different results and conclusions can be obtained by different analysts attempting the same social media data search. In extreme cases, analysts can create such highly divergent queries that the associated data leads to different answers to even simple questions, such as: Which brand is my main competitor? Is Product1 more of my brand's conversation this month centered around product? Is the sentiment expressed toward my brand this month more or less positive than the sentiment expressed toward my brand last month?