Digging for Likes: Is Social Media Mining the Future of Market Research?
You can find out a lot about someone just by looking at their various social media profiles. From careers to vacations, circles of friends to book recommendations, it seems like there’s a social site for just about every aspect of life. So it makes sense that market research, as an industry, would look toward social analytics as a method for assembling accurate customer profiles.
But it’s not as foolproof as you’d think. While it’s true that most social media users have a habit of oversharing, simply mining the data from active users might not provide the accuracy necessary for use in market research results. Understanding the minefields of social media mining makes the difference in how you gather–and use–social analytics.
Here’s the most glaring issue with social media mining and its results: The average person who uses social media may not be the average person in general. When piecing together a customer profile, those who share the most won’t necessarily represent the population at large. Chances are that it’s a higher-profile person, and while you can glean much from their personal accounts, it’s a case of too much information about those you don’t need, and not enough information about those you do need.
Data mining can also struggle with a lack of test datasets. Unless you’ve been mining social media for an extended period of time, it’s difficult to compare and cross-validate results for the most complete profile. It leaves data miners wondering just how much of the information skimmed from social media for market research is actually relevant from a market research point of view.
Hitting Pay Dirt
The trick to social media mining for market research is utilizing the right tools to dig deeper. You can’t take likes, shares, and tweets at face value: Instead, the nuggets of wisdom are found within the context.
Topic “bigness,” for instance, can make a huge difference in results. Before mining social media to find out a person’s attitude toward a brand, the topic must be big enough to generate interest and conversation to create a wider sample. The solitary person who rants about a brand online is not indicative of the general population, so social media mining works best when a topic is widespread and trending.
It’s also possible to thicken up what may be fairly thin data. A Facebook profile, on its own, isn’t groundbreaking. But when an individual’s Facebook page is contrasted to his or her Twitter and Instagram feed, forum posts, and other social networks, it creates a more complete customer profile. By identifying the same user across multiple platforms, you get a better sense of opinions, likes, and habits.
Social media mining isn’t perfect, and it probably won’t replace any market research company’s strategies that have been working until now. Still, the very nature of social media–vast, opinionated, honest, and behavior-based–makes it the ideal add-on for creating customer profiles or sampling attitudes toward brands. In the end, a little extra digging through social networks could yield a big pay day.