How To Get Deep Insights About Customers
From Your Social Media

by Ian D. Parkman, Ph.D. and Sam Holloway, Ph.D.

Marketing Intelligence Module
Topic 3: Social Media

Picture a single room filled with all of your customers and potential customers; off to one side a woman with a dog is holding court in the center of a huge group of rowdy merrymakers, while on the other side a tall guy in a green hat is quietly whispering with a few close friends, and right in the middle of the room a balding man in a kilt is jumping up and down shouting to the rafters, but no one is listening. This room is our analogy for the world of social media. It’s loud, it’s busy, and all of these characters have an opinion to share about you and your beers, the question is: What are they saying and should we be listening?

While most, if not all, craft beer brands appreciate the importance of using social media to connect with consumers and to build brand positioning; what, really, is the strategic and competitive value of simply counting views, mentions, followers, favorites, or likes? Most craft breweries have a presence on Twitter, Facebook, Foursquare, Untappd, and Instagram, but the metrics available to evaluate the time and money spent on social media are broad and imprecise. Without any sense of the valiance—that is, the positive or negative sentiments expressed by the consumer, who is listening, or the actual content of those interactions; how can you differentiate the loud and listened-to woman with the dog from the merely loud bald man in the kilt? And how do you recognize and leverage the value of the seemingly influential guy in the green hat?

The purpose of this white paper is to provide busy brewery managers and their marketing teams with a simple framework and some free tools to move beyond superficial summaries of their social media accounts (and it is yours, by the way) to draw deep and actionable insights based on sound research principles and data.

Social media platforms offer all brands, from Fortune 500 pharmaceutical companies to local family run breweries like Backwoods Brewing (Carson, WA), a powerful opportunity to gain instantaneous access to the unfettered opinions of consumers.  Social media allows a tantalizing mix of information, but it can also suck away productive time when you should be making beer, paying bills, or (more likely) paying attention to your spouse or children after a long day of making beer!  Social media allows you to know who is talking about your beer right now, what are they saying, who are they with, what are those folks saying, where are they saying it, and who is listening? Contemporary content analysis of social media data allows you to listen in. This is a truly amazing advancement in the field of marketing. Imagine that only a few years ago even the largest and most sophisticated companies in the world couldn't access these seemingly simple data-- Now its everywhere!

We want you to focus in on only two concepts that allow deep insights from social media data: Social Media Analytics and Content Analysis. Content analysis refers to a general set of techniques useful for analyzing and understanding collections of qualitative text, such as the reams and reams of content provided through social media tweets, tags, and updates (in contrast to quantitative measures of things, such as sales data). There is considerable work done in this area, which predates social media research by decades. However, newly available content analysis software allows easy word counts, reach analytics, word clouds, volume, sentiment analysis, and influence measures for valuable, up-to-the-minute snapshots of social media content. To make sense of this world, a few easy ground rules may be helpful. The following are the important steps we’ve observed for analyzing social media data.

Step 1: Developing a problem definition and objectives

For most social media research projects, especially research conducted by smaller craft breweries that don’t have the time or money to waste spinning their wheels—the most important step is to focus on a single, specific, measurable, definition of the business problem you hope to investigate: What are we trying to do? What decisions will be made with the information we collect? What are the boundary conditions that we are working with? ‘Boundary conditions’ is an academic term asking if you know, as specifically as possible, the limits of the question you ask—e.g., “I wonder why our Winter Seasonal got so much more positive attention during the week of Thanksgiving than normal”, “We think that left-handed women seem to be our best customers”, or “Our closest competitor saw a big increase in visibility after they sponsored that local Oktoberfest”.  Each of these research questions helps to narrow down the universe of data that you’ll be looking at so you can start looking for patterns. Once you have a research objective nailed down, you have a better idea of what you’re looking for among your social media data. The alternative to a clearly developed research question is simply wading into your Twitter or Instagram feed looking for trends. An exercise we cannot recommend unless you plan on being very lucky, or very good.

The following are examples of objectives that are particularly well suited for social media analysis:

  • Service quality; Competitive analysis—e.g., the tactics and strategies of your competition

  • The performance of product extensions

  • Product strengths and weaknesses

  • New uses for your products

  • Consumer targeting—e.g., tracking reactions to specific advertising and promotions in geographic areas

Step 2: Identify key search terms

The identification of appropriate search terms is a crucial step to the successful analysis of social media data. Traditional content analysis calls this ‘coding’. What you’re doing is creating a list of 3-4 key phrases, terms, or variables that may be of interest to your objective.

  • For example, many of our clients are interested in how the successes (or failures) of a local sports team may relate to their brands. Accordingly, search terms—i.e., codes-- such as “#Packers”, “#GBvsSF”, “#Monday Night”, #MNF”, as well as more idiosyncratic terms such as “#couch”, “#big screen”, “#winning”, or “#losing”, and “#sadness” may be revealing of clear patterns of behaviors among your brand community

However, keep in mind that the process is often an iterative one, with broader searches being followed by searches using combinations of terms or newly discovered synonyms or tangential phrases (e.g., “#hand-made” leading you to “#artisanal” as a relevant term). Trial and error is common and is to be completely expected.

  • Obvious terms to start a search include your brand name, competitors’ brand names, your product category, local bars or pubs, and neighborhoods in your area.

  • More exploratory analyses might investigate specific events, feelings or emotions related to a brand. For example, a stretch of cloudy, rainy weather in the Pacific Northwest of the United States may engender a flurry of social media activity by consumers settling into a cozy spot by a fireplace to enjoy a warming espresso stout. Terms such as “#cozy”, “#comfy”, “#snug”, or “#cuddle” may be useful to reveal a pattern to how your consumers are feeling at the moment.

Step 3: Identify social media data sources

Once you have your research objectives and your search terms; you’re ready to start digging. The first step of this stage is to identify data sources.  Online aggregator tools, such as TweetDeck and Scout Labs (now part of Lithium), can aid in this process. However, keep in mind that these tools simply provide broad “counts” of your content, its up to you to interpret your results and provide context.

Secondly, depending on the research objectives, you may want to categorize your followers in some way—i.e., create some way to distinguish the isolated ranting’s of the balding guy in the kilt from the legitimate and more potentially dangerous complaints of the woman with the dog. Some types of social media sites that can provide sorting of consumer-generated data include the following:

  • TweetReach (http://tweetreach.com/) and Klout (https://klout.com) are useful for measuring the impact of your social media, guiding you toward your most influential follows and the content that engenders the liveliest discussions

  • Social Mention (http://socialmention.com/) is a free analysis tool that measures the influence of your keywords, hash tags, and users on four categories: Strength, Sentiment, Passion, and Reach—we particularly like these themes because they are extremely useful for drilling down and making sense of your social media traffic

  • Simply Measured (http://simplymeasured.com/) provides a variety of free reports analyzing your brand’s traffic, engagement, trends, and influencers

  • Google Analytics (http://www.google.com/analytics/) is still the best one-stop shop for a comprehensive web analytics tool for measuring social media traffic at the individual link-level, real-time traffic, and conversation tracking—all summarized in a handy and easy-to-interpret dashboard.

  • Zuum (http://www.zuumsocial.com/) is pay site used for benchmarking performance relative competitors, tracking the content that other brands post, and identifying their key influencers that you may want to engage with

Step 4: Analyze and Interpreting your data, and then Strategize

After you have gathered and organized the social media data related to your research question, we recommend a few best practices for analyzing and interpreting some useful outcomes. First, you should review the data thoroughly. And then review it again. And then have someone else in your team review it. As with all research, insightful analysis depends on a comprehensive knowledge and understanding of these data. Any actions you take based on what you find have to be grounded in your complete confidence in the quality of the data and deep understanding of your product marketplace. Your objective is to find evidence in your social media data that addresses your research questions. More specifically, you’re looking for key themes, beliefs, ideas, concepts, definitions, or behaviors from your streams—as in the cloudy day example from above. In systematic qualitative research experts use software (e.g., NVivo) to provide ex nihilo (“out of nothing”) codes, maps, and groupings of the themes in the data. While this provides the benefit of unbiased results (e.g. letting the data speak for itself), the software involved is expensive, time-consuming to learn and operate, and probably far beyond the needs of anyone reading this white paper. We suggest that, far from striving for unbiased results, embrace your subjectivity! No one knows your products, customers, and competitors better than you do. Remember, all you’re trying to do is find patterns in the data to help your decision-making: It seems like my consumers think “ __________” based on their social media traffic is already better than making decisions based on your gut instinct.

Once you’ve gone through your data looking for patterns, you should synthesize your thoughts in a brief research report. Start by describing your research question and why you feel it demands an answer before using concrete examples and illustrations from your data to clarify the question. Here is where social media data really stands out; quotes can be presented from Twitter, images can be displayed from Instagram, and reviews from Foursquare or Untappd can be used to precisely illustrate, in a consumer’s own words, their thoughts, feels, emptions, and opinions. This data is gold. With focus groups, interviews and even online communities, consumers are responding to directed questions. The users of social media provide you a glimpse into whatever is on their minds-- and oftentimes provide the time and locations where those thoughts occurred. This represents a great opportunity to gain new understandings about their motivations and create dynamic interactions with consumers. For example, imagine the potential to connect with consumers through social media by commiserating with a local sport teams loss, basking in the glow of victory, or mirroring back your brand’s desire to open a barrel-aged stout on a stormy afternoon. These may appear to be trivial actions, but as we’ve discussed in our Building Brand Community sessions, connecting with the zeitgeist of your consumers is how brand communities are formed.