Social Media: The Next Generation of Market Research Is Here

Posted on December 10, 2013


“Social analytics tools will become the must-have to gain insight, make better, faster business decisions and improve customer satisfaction.” – Social Business News

Steve Jobs was famous for taking a stand against traditional market research, telling us, “People don’t know what they want until you show it to them. That’s why I never rely on market research.”

Click to Tweet This Post
★ Social Media: The Next Generation of Market Research Is Here

But as Bob Gilbreath pointed out in his article, The Steve Jobs Market Research Quote Should Rest in Peace, “One should never conduct a research study that asks people what they want…First,  you must spend time understanding and gaining insights into consumers’ existing habits, beliefs, routines and unmet needs.”

Get the latest news and insights for senior marketing pros by subscribing for free to The Steveology Blog by email.
*Privacy Policy

This renewed approach to market research – understanding customers in their natural environments rather than asking them what they want – is exactly why social media listening is such an exciting opportunity for marketers today.

Focus groups, surveys and user interviews all have their places, but social media research is quickly emerging as a faster, more cost effective way to understand what people are already talking about on their own. This information can be used to help us as marketers to better understand how our customers feel about our brands and our competitors, to develop better products, to uncover new content ideas and to improve the creative initiatives that we’re already putting out into the world.

Every single day, there are 175 million new tweets posted online, along with two million blog posts, 250 million new photos and 864 thousand hours of video. This content is rich with actionable insights for those marketers who are wise enough to uncover trends and qualitative context within the posts.

Where Do We Start?

Any marketer who has spent time using automated social media analysis tools knows that these tools can be incredible frustrating. Automated sentiment is still only around 60% accurate, and in my experience, anywhere from 5-70% of keyword search results around any given brand are typically either irrelevant or driven by spambots (particularly in the retail word).

At the same time, it’s impossible for any human analyst to go through and sort all of the relevant posts that are being shared about any given company or brand each day.

So where do we go from here?

I recommend a balance between the human and the machine. If you’re evaluating more social media posts than a human can possibly get through, use a representative sample set to get a deeper level of accuracy and context from these conversations. Go to and enter the total number of search results in your population to determine how many posts you will need to evaluate in your sample set.

You will need to understand two basic concepts to create your sample set:

  1. Confidence Level: The confidence level will tell you how close you can expect to be to the true number based on your sample set. We typically use a confidence level of 95%, which means that we are 95% certain.
  2. Confidence Interval: The confidence interval, or margin of error, will tell you the sampling error of your results. We typically use a confidence interval of five, so if we find that 50% of social posts are being driven by a specific brand initiative, we are 95% confident that between 45% (50-5) and 55% (50+5) of all social posts from the entire population are being driven by that brand initiative.

You can use Excel to generate a completely random sample of social posts, or you can simply divide your total population by the number of posts that will need to be analyzed (the number of posts that came through in your search results), and tag posts at that interval.

For example, if you have 4,000 total posts to analyze with a sample size of 351, tag every 11th post in your set for sentiment and context. You’ll typically need to tag just under 400 posts to get a meaningful sample size.

Once you’ve added context tags to categorize your posts, you’ll be able to gain a better understanding of what topics are driving the most conversations around your brand, competitors, industry or other topic of interest. Use this information to answer these questions:

  • What do our customers or potential customers think about us?
  • What do our customers or potential customers say about our competitors or industry?
  • What questions do our customers or potential customers have about our products or services?
  • Where and when do they talk about us and our competitors?
  • What drives positive conversations around our brand or our competitors’ brands?
  • What drives negative conversations around our brand or our competitors’ brands?
  • What topics or attributes drive the most engagement around our brand or industry?

Listen Before You Act

Forget about figuring out the hottest new social media tactics to promote your brand. Before you take another step in your marketing plans, make sure you’re listening first. Marketers who can harness the power of listening to their customers in social media will be leaps and bounds ahead of their competitors.

About The Author

Allie Siarto is the co-founder of Loudpixel, a social media research company that helps companies better understand their customers and markets by analyzing social media conversations that are already taking place each day. Loudpixel has worked with global Fortune 500 companies and agencies across the technology, restaurant, finance, auto, retail, consumer goods and healthcare sectors. Allie’s work has been published in Forbes, PC World, The Washington Post, Upstart Business Journal, Yahoo! Small Business Advisor, MSN, VentureBeat and Under 30 CEO. She is the author of the book  The Social Current: Monitoring and Analyzing Conversations in Social Media and teaches a class on the same topic at Michigan State University.