Listen & Learn: Monitoring Social Media for Actionable Business Insights
1. Listen & Learn Monitoring Social Media to Draw Actionable Customer Insights
2. Hello. @KatFrench Social Media Manager Google Analytics IQ Inbound Mktg Cert. Pro. Featured contributor, Social Media Explorer AdAge 150 blog LinkedIn: www.linkedin.com/in/katfrench kfrench@doeanderson.com 2
15. The 4 Types of Social Data 15 “too hard” “too soft” “just right” “sweeter” “saltier” “faster” “more accurate” “best” “worst” How much?How many? Peripheral Random Unexpected Surprising “Off Topic” Like Dislike Love Hate Prefer Avoid
20. The How and The Why of Drawing Insights Pull the unusual and significant data. Look at it across the 4 types of data. Apply context from your organic business knowledge. What action does the story suggest? Marketing and Advertising Insights Sales and Support Insights Innovation and R&D Insights 20
21. Top Tips for Activating Insights Across Your Business Don’t sit on your insights. Don’t create data-dumps. Do have a list of activators. Do give appropriate lead time. Don’t create rainbows & unicorns. Don’t create tempests in teapots. 21
Which came first, the chicken or the egg? Will we ever know? Probably not. What we do know is that you need BOTH to complete the life cycle. What does that have to do with social media? Think of the egg as “listening” and the chicken as “activation.”
Jeremiah Owyang of Altimeter released an excellent graph that illustrates social CRM, which is a holistic, complete lifecycle for corporate social media. You’ll notice that the listening/insights use cases always come BEFORE the activation uses.
While a lot of lip service has been paid to “listen first!” most of the attention has been paid to activating social media. “Listen first” often ends up meaning “Listen until you overcome your fear of engaging.” Listening is not an egg that you throw away once the chick of activation has come out of its shell. The whole lifecycle is cyclical and constantly repeated. Listening shouldn’t stop as a separate activity once engagement starts, but it often does.
When listening is no longer a separate activity from engagement, it changes the way you listen. You’re listening for the next engagement opening. Or you’re listening for the next potential customer service fire to stomp on. You develop tunnel hearing. Potential insights become instant “to dos” or they’re disregarded. If it’s not immediately actionable, it’s discarded. Without a separate responsibility for listening and drawing insights, activation becomes mere reaction, fire-stomping and slightly creepy sales maneuvers.
Let’s look at this another way. This chart comes from Chuck Hemann. Notice what takes up the big section in the middle? Distilling signal from noise and relaying that signal to the appropriate activation team. Have you ever seen this actually happen? “Signal” does not mean “immediately actionable.” It means “meaningful.” Sometimes, that’s “meaningful over time.”
In looking at the previous two graphics, it’s clear that the real missed opportunity in social media monitoring is insights. So in the context of social media monitoring, what are insights?
First, what insights are not:Data dumps are not insights. In fact, a data dump is the opposite of drawing insights because aggregate, unprocessed data obscures meaning. Pretty charts and graphs are also not insights. We all love a good infographic (the ones in my slide here are from xkcd and informationisbeautiful). Some of us genuinely are visual learners. However, charts and graphs can really only show what numbers mean in relation to other numbers. They don’t necessarily imply a real-world application, and by themselves they don’t provide any narrative, or story context.
Along similar lines, tools don’t provide insights. I love social media monitoring and social analytics tools, but the vendors are overstating their case when they claim to provide insights into your customer. They provide data from which a person can draw insights. Tools can collect, collate and compare data. They can’t create context. They can’t provide information that’s actionable at anything other than a knee-jerk, “someone said our company sucks!” level. That is why you need people to support listening.
Sadly, this is not a photo of the distillery with which I am most familiar. They have only one still, and it’s too pretty to cover up with text. However, I want to make the point that insights are drawn and distilled from your data. Which means you have to be immersed in that data. I could take this opportunity to tell the joke about the man who fell into the beer vat and drowned, but only after they fished him out twice, but that’s been done before. So we’ll just go on to say that you can’t really draw insights unless you’re spending a sufficient amount of time really diving into the data to the point that you know what’s “normal” and what’s new and noteworthy. Again, this is another reason why people are important. People provide continuity.
Once we’ve distilled the new and noteworthy out of the aggregate data, we have the material for pretty charts and graphs. But as we noted earlier, those only give the meaning of numbers in relation to other numbers. When we apply context and implication to those numbers, they tell a story. Look at the tower, with a lone window. When we apply context to the image “tall, forbidding tower with no access” we imply “prisoner.” Or possibly “watchtower” and “guardian.” The more context we have, the more detailed and accurate the implied story becomes.
… but it’s a reasoned, deliberate action. Else, we’re back to a reactive view of listening. If we go back to our prisoner in a tower, what action does that story prompt? A rescue attempt? Or if the prisoner is a terrible criminal, should we fortify the tower? Again, context is key.I know I may have lost some of you when we went from business data to princesses in towers. But I don’t know your business. I’m not swimming the fermentation tanks where conversations around your brand happen. I’m going for a common vernacular here. We all know story. We all know a story with an invitation to act at the end. Very often, social monitoring and measurement gets assigned to people who see numbers as an end to themselves. I’m trying to get you to see your data in a different way, as source material for a true story you understand clearly and know how to act upon.
So now that we know that insights are the stories about your brand that are distilled from data, understood by people, and which beg a response, let’s look at the process of getting from data to actionable insights. We start with data, but let me say that I’m not going to go into any detail about specific monitoring tools that provide this source data. Partly because I’ve worked with three or four of them and they all do an acceptable job. Mostly because your preference of tool is going to be highly personal and subjective, so the best way to find a tool is to test drive all that are in your price range. I also don’t want us to get hung up dissecting which tool is betterand miss the real story, which is “what you do with the data.” If insights are distilled from data, we need to look at the types of data that most social monitoring tools can provide. There are four basic types of data we have access to: Quantitative,Sentiment, Qualitative, and Subjective/Narrative. The different tools measure things differently, but if your tool can provide you numbers of mentions, positive or negative sentiment for those mentions, and allows you search or sort within your topics for descriptive phrases and frequently-appearing words or phrases, then you have access to those four basic types of information.
So let’s start with the simplest to understand: quantitative data. And from there, we’ll see how all four types of data inform each other to create meaning and tell a story you understand and know how to act upon. In order to draw meaning from your data, we need to know “what’s the baseline?” What constitutes “situation normal” for your business, and what’s unusual. For the stuff that’s unusual, we then need to know if it’s significant or just a random blip.This is why quantitative data matters. It helps us find “how much” is normal (establishing the baseline), and “how much” of the unusual stuff is out there (is it significant?). If there are typically 40 mentions a month about your mocha latte, and suddenly you’re seeing 80 a month, and the new word “nutmeg” popping up in several of them, then maybe one of your baristas is getting creative with the spice rack. Or maybe a few of your customers have been adding a dash of nutmeg to the mocha latte, and started a trend.In the context of this business example, a sudden spike in quantitative data (post volume) is your tip off that something’s out of the ordinary. Digging deeper into the subjective data, “nutmeg” is an unexpected term, and multiple mentions make it significant in your particular business case. You’ll want to drill down to sentiment on those posts and see if the buzz on the new nutmeg mocha is good or bad.
Which brings us to sentiment. Clearly, the lady attached to that foot has had an extreme change in sentiment towards the gentleman dangling off the cliff.As with the volume of mentions (your quantitative data), it’s important to establish a baseline measurement of sentiment. There are some companies and industries that people just love to hate. Don’t assume the goal is 100% positive sentiment—that’s not realistic. Once you have a baseline, and you’re checking with sufficient frequency, you can spot a sudden turn in sentiment. Or in the case of our spicy barista, some change in volume, qualitative or subjective data might prompt you to check the sentiment of that particular conversation subset.The action prompted by a turn in sentiment is investigation. WHY is sentiment suddenly moving more positive or more negative? Often the answer to that question is in your QUALITATIVE data.
Who is “top dog” in your competitive set? What’s the criteria? If you’re noticing a significant change in sentiment about your product or brand, what qualitative language do you see in conjunction with that? Do people say you’re better than your competitors? Faster? More convenient? Is your product saltier, more durable, cheaper, more fun? Who in your organization could use this insight? Can marketing use it to improve the accuracy of your positioning (social marketing insights use case)? Can sales use social testimonials in their literature? Would those recommendations close more deals for business development? (social sales insights, proactive lead generation). Now you see we’re getting to the “actionable insights” piece. With three different sets of data, surrounded and infused with context from knowing your business, you have a story you can act on!
The last type of social data is what I call subjective, or narrative, data simply because I can’t think of a better name for it. Subjective or narrative data is simply the oral tradition of your product or brand story, as it’s retold by people on the social web. Again, you have a baseline, which is your brand story, and your advertising messaging. So the subjective/narrative data can tell you how effective your advertising is, when you hear that story repeated back to you from customers. That’s indicative that your advertising is resonating with people. We’ve also used narrative data to determine when ad messaging wasn’t ringing true. A computer can’t detect sarcasm, but an analyst who’s immersed in the web culture can. Which is another reason it’s so important to devote people to this, not just tools. What’s outside that baseline story, though, and significant, represents new opportunities. This is where the innovation insights and crowdsourced R&D come into play. Sometimes your customers find a new way to use your product. Or they ask for a feature you couldn’t have predicted they’d want. Those are actionable insights that can shape the future of your business and provide new opportunities. Going back to our coffee house illustration, “nutmeg” is the surprise factor in your recent spike in social buzz. It’s the twist in the story you didn’t see coming.
So now you’ve seen a simplified example of how the four types of data inform each other. You’ve seen how:- pulling the unusual and significant in your data, - looking at it from the multiple perspectives of the different types of data you have available, and- applying context from what you already know about your business, you can draw insights that can be used to improve your business in a variety of ways. Insights can help you redirect your advertising messaging when it’s off-target, or encourage you to stay the course when it’s on-target and resonating. They can help you find new opportunities for innovation in your product and service offerings. They can warn you when there’s a sea-change in your industry just beginning. These kinds of insights represent the vast and largely undertapped potential of social media listening.
So, before we dive into any Q&A, I thought I’d finish up with a quick hit list of tips for activating insights across your business. These are taken from some of the common missteps I’ve made or observed in the last year or two as social media monitoring has matured. Thanks so much for your time, I hope it’s been helpful.