Market research is basically an organized effort to define markets and gather information about them and the consumers they represent. Online consumer conversations and social media analytics are a natural fit for market research. Yet, many organizations view them simply as a way to measure the success of social media marketing efforts and campaign performance.
But this relatively narrow use of social media analytics presents a major drawback. Campaign analysis using social data only works if the consumers on social networks are talking about a brand’s marketing activities. So, it’s only effective for industries or brands where there’s a built-in affinity, where it’s top of mind.
Think of politics — conversations about the president. Or what TV shows are trending. Events like the Super Bowl or the Oscars. Or the latest iPhone. People talk about these things. But they’re not talking about their credit card or their insurance policy, unless there’s a big interest rate or premium hike. They’re usually not talking about the sponsorship their bank did at the U.S. Open tennis tournament. Though savvy brands can measure the impact of sponsorship investments using image analytics.
Social Media Analytics for Market Research
Social media analytics for market research offers a great deal of broad-based insight. It allows analysts to “eavesdrop” on consumers as they go about their daily lives, revealing biases, opinions, preferences and passions that may ultimately affect behavior.
San Francisco Chronicle columnist Leah Garchick has run a public eavesdropping feature in her column for more than a decade. These are things her readers overhear at the coffee shop, the art opening, the dentist’s waiting room, the hotel lobby. You get the picture. Imagine that times a few million… or more. Of course, just because social listening conversations are unsolicited doesn’t mean they’re always valuable. They could be astonishing game changers, status quo confirmations, or random noise.
One of the keys to a higher percentage of the former is starting with the right questions.
Establishing a Research Process
No matter how powerful your analytical tools, they won’t deliver usable insights if you’re not asking the right questions. It’s like having a classic Fender electric guitar. If you want music instead of noise, you’ve got to learn how to play. “Learning how to play” in terms of analytics is where the research process comes in.
An often-overlooked factor that contributes to developing a research process is the general level of understanding within an organization about what social media data is and what it can do. There’s frequently a knowledge gap between the analysts who do the research and the executives and line-of-business managers who want to use consumer insights from social media to inform strategy and tactics.
Closely related to an organization’s understanding of social media data is its data sophistication. Do researchers and analysts understand the nuances between structured, quantitative data and unstructured, qualitative data? And that social media data are completely subjective and open to interpretation? Both types of data are important.
This may be where competency in statistics comes into play. For example, interpreting sentiment from social data may need a statistical approach that employs machine learning, where it’s not so much about being more right. It’s trying to be less wrong. In computational learning theory, this is known as probably approximately correct (PAC) learning (Leslie Valiant, 1984).
You’re trying to train a model to recognize something based on attributes that you think may describe it. This is different than an analysis using quantitative data that produces one result. What the result means may be open to interpretation, but if the data stays the same the result doesn’t change.
A Market Research Cheat Sheet
Some guidelines for the market research process include:
- Know what you’re looking for. That means narrow the scope of research, so it addresses a specific business question or goal. Something that either can’t be answered by other means or, possibly, to confirm results from other sources, such as a survey. Good research always prompts more questions.
- Think strategically, not tactically. You can’t reliably tie social media metrics directly to business goals. Just because you generated X-million impressions during a campaign using social media, doesn’t mean that you can attribute an increase in sales to that alone. Correlation is not causation.
- Eliminate ambiguity. A question such as, “What is trending in the consumer electronics space?” is too broad to be useful. Better would be, “What does the current state of VR technology mean for companies that make processors?”
- Explore nuance. Take a simple example: Black Friday. Every year someone predicts the death of Black Friday. But to paraphrase Mark Twain, reports of Black Friday’s death have been greatly exaggerated. But Black Friday has certainly changed over the last decade. The nature of those changes is something that could be explored with online data. And understanding those changes could be very helpful to marketers.
- Think like a journalist. Whatever the topic, search for the interesting angle or perspective. Not so much is something this or that, but if it’s “this,” then what does that mean? Why is this phenomenon occuring? Sometimes, one Tweet or anecdote prompts another question and triggers a research project that reveals an insight no other data source could have possible uncovered. Initial ambiguity can be a clue to something worth exploring.
The Quantitative Side of Qualitative Data
Sometimes, there’s an organizational “queasiness” about social data because it isn’t “hard” research. But the volume of social data can easily bridge one of the most common gaps in research, which is sample size. For example, Crimson Hexagon has 100-percent coverage of Twitter and Tumblr. So, with access to all that data, our customers benefit from statistically significant samples.
When you have access to an entire dataset, you can be more comfortable generalizing about what the data says — whether a trend is growing or shrinking. And, you can quantify the data in several ways. X percent of social conversation about a topic skews positive or negative. That’s a number.
Netflix might want to know this about its original content. Which shows are generating the most conversation? Again, that’s a quantitative view of qualitative information. Combined with other data Netflix has, such as how many people view a show, how many skip episodes, what is the repeat rate for shows or episodes and so on, a clearer picture emerges of what customers think.
Get consumer insights delivered to your inbox.
An Outsider’s Perspective
In the right hands, social media data is a rich source of consumer insights. For many brands, an outside agency or a dedicated in-house consumer or market-trends SWAT team can help them learn how to think about social data and its place in market research. Regardless of who analyzes your online data, executives and business-line managers need to understand its value. The consumer insights it offers are too important to ignore.
To see more examples of how consumer insights from online and offline data sources are helping businesses to compete and innovate, check out our business insights guide.