Social media research, customer insight, and the power of the one

meaningless wordcloud

I’d like to tell you a short, cautionary story today to explain why I am wary of “word clouds” and most other trendy displays of social media data and “sentiment.” By the end, I hope you will be wary, too.

Long before the days of Sysomos, Radian6, or even Facebook, the only way to learn about trends and “sentiment” was to actually poll your customers or go out and talk to them.

When I was a marketing director for a large company I led an activity called “Listen to the Customer.” To augment regular customer surveys, every other year we would get on a plane and actually visit with a cross section of customers face-to-face. This was not a “nice to see you” visit. This was a rigorous process to help identify trends, ideas, and ways to improve our product.

One year, after we had spent weeks on the road, we concluded our trip with a group of researchers at their headquarters in Atlanta. In the very last hour of the very last meeting, one scientist mentioned in an off-handed remark that he had seen a new preliminary government study indicating that a chemical in our packaging might be related to health problems.

We were stunned. WHAT?

Although the research was indeed preliminary, if it turned out to be true this would be a devastating development that could bring our company to its knees.

The power of “the one”

As a cautionary measure, we immediately began a multi-million dollar effort to eliminate the suspicious chemical from our product. This was a very unpopular program both internally and externally because there was no proof of a problem yet. After three years of research and customer testing, we introduced new packaging that was completely free of the chemical.

Two years after we had made the change (five years after that customer meeting!), a front page article in the Wall Street Journal announced conclusive new research linking the chemical to a variety of human health issues. We were in good shape. Our competitors were not.

I wanted to tell this story because if we had depended on a word cloud or a chart of “sentiment analysis” to tell us what was going on with our customers my company would have never made this discovery and avoided a catastrophe.

These new analytic tools can make us lazy marketers. In the past few weeks I have sat in meeting after meeting looking at customer “research” based on word clouds and beautifully-designed infographics that tell us NOTHING about what is really going on in the marketplace.

Dig deep for insights

We will rarely get break-through insights from a high level data summary … which is probably the same thing our competitors are seeing any way. Insights that lead to innovation do not normally come to us through lists of averages and word counts.

Do you really want to find competitive advantage? Get out of your word cloud and spend a couple days with your customers. Ask them what they love. Ask them what they hate. Ask them what keeps them up at night. Dig deep. Ask why, why, why, why, why.

Do. The. Work.

What is your take on this issue? I’d love to hear from you in the comment section!

All posts

  • I ran a word cloud on this article and the biggest word was “why.” I think that’s appropriate.

    This is a soapbox for me, so I’ll refrain from hijacking your comments section. But I’ve spent a lot of time examining “social media research.” Aggregate data like the kind you mention is this article will rarely, if ever, tell you much. Do the work, indeed.

  • Pingback: Social media research, customer insight, and th...()

  • I had a hunch this would resonate with you! Thanks for dropping by.

  • In the Lean community, there is this saying: “Go to the gemba” (because all cool Lean terms have to be in Japanese – well, gemba means the place where the work is actually performed).

    Funny historical anecdote, this belief that one should always go and see for oneself was partially a reaction against management by reports that had been built at GM in the early 20th century.

    I find the same sentiment in your post, and I heartily agree.

  • PeterJ42

    Soapbox for me too, Tom. In the connected world, it has never been easier to get comments from customers, to correlate these into quantitative data and to build a 2-way connection with customers. But few companies even try – they hide behind “What do you think” on their blog, no-reply on their email and a disconnect between sales and company which means comments made to sales people never reach the people in the company making decisions on product or service.

    One other question, Mark. We now live in a world where product lifecycles are measured in months and company business models change every few years. Stories break and kill companies in minutes. How on earth did it take 3 years just to change a piece of packaging? Do you think you’d get away with this today?

  • I love that. I would like to be the blogging gemba : )

    Many thanks for your wise observation Ville!

  • It was a very complicated and risky change that involved many iterations and years of shelf-life tests to make sure there were no compromises to quality. It didn’t help things that customers did not initially share our sense of urgency for change! Sometimes, there are no shortcuts.

    Thanks so much for commenting Peter.

  • Spending time with your customers and constantly asking “why” is always time well spent. 🙂

  • PeterJ42

    I don’t want to sound critical Mark – I’m on your side and this is a good and sensible article – it needed to be said.

    But it does seem companies have become anti-social. They are over-thinking things. And that this is leading to analysis paralysis.

    This is a great example. It is packaging, after all, not what the customer is actually buying. If we really cared about the customer, we’d have it off the shelf day 1.

    That’s reputation protection too – if that WSJ report had come out before you’d changed along with a leaked email that you knew about it 3 years ago the company may not have survived, even if you’d done more than competitors.

    But we digress from your core point. Company staff – in many departments, not just marketing – have become expert at finding ways to show other managers that they understand the customer – but these are actually several steps removed from what customers actually think.

    What are they afraid of – that customers will tell you something you don’t like – or that makes you look bad to your boss?

  • My favourite question: “Why?” Cheers! Kaarina

  • Hey Mark-

    Great work as always, but I don’t think the two are mutually exclusive. The social team might create word clouds to send to the communications and PR teams, but our technical and product development folks very much get on planes and see customers. I don’t think you should renounce the word cloud because it only gives you top-level views, but use it for what it’s intended for – top-level views! 😀

  • On my, I just love this!Mark, you’re explaining the reality. By the time the market is aware of a situation (ie: becomes mainstream), it’s too late. Marketing needs to be as much about “seeing what’s coming” as it is about “knowing what’s here”. IMO, what’s most interesting about this is that SM Analytics is very much about both. In fact, if users really think about the “early warning” capabilities, it’s pretty powerful stuff. The problem is to many “marketers” (and their employers) are pressured to think about the here and now (as that is how they are measured in the near term) and not planning for the future.

    But as you clearly stated, SMA is only one part of a very complex puzzle. Always has been, always will be.

  • +1 Laura : )

  • Mine too. It solves a LOT of problems!

  • I renounce it only because it makes marketers lazy. Something like a word cloud is VERY limited its use. Your social team and marketers should go out to see customers too. In fact, if at least 75% of your meetings are not with marketers something is wrong.

    Top level views have a place of course but rarely create insights that lead to change and improvement and that is what we need to be pushing for every day! Thanks so much for your comment and dissent Eric!

  • i knew this would resonate with you and I love the added spin you put on here Steve. “What is” is ever so less interesting than “what will be!”

  • Actually, the package WAS what the customer was buying. It was food packaging in direct contact with the stuff that people swallow. For this reason, years of shelf life data and intense scrutiny were needed and appropriate. A wrong decision could literally put a brand in jeopardy.

    Although some expedited testing does occur in the food industry, generally lengthy testing and approvals are needed before any new type of packaging reaches consumers, and this is a very good thing!

  • Always love reading your take on social media research Mark! Learning keeps me going… Thanks for sharing your passion.

  • My pleasure. Glad it helped!

  • Claudia Licher

    Love this article Mark.
    It’s true. Word clouds, infographics and other visuals are great for presenting information, but they won’t tell us anything we don’t already know (or could know)… it’s up to us to go out and discover the stuff we don’t know anything about.

  • I just hope the new generation of marketers remembers that!

  • Claudia Licher

    If they don’t do so right now, they will learn the hard way I guess.
    Just realized that “The stuff we don’t know anything about.” had a “final frontier” kind of ring to it. Or perhaps it’s just time to head home for today.

  • Patricia Haag

    It’s amazing just how much research is necessary to find out exactly what people want.

    I’m impressed that your company went forward with the change in packaging based on an off-hand remark about a preliminary study. Amazing to see a company look past the next quarter’s revenue.

  • This is true!

  • We knew it was the right decision but had to fight through the quarterly numbers mentality.

  • I couldn’t agree with you more, Mark. Too often companies will depend solely on other people’s research. I’m not saying to discount third-party market research, as there’s definitely value in that. Companies should also ask their customers for direct feedback, as this is a surefire way to hear it from the proverbial “horse’s mouth.”

  • And with the pace of change these days, we need to do that more often! Thanks for commenting Giovanni!

  • Scott Valentine

    The Heisenberg’s Uncertainty Principle basically states that the simple act of measuring something changes it.

    In my experience, customers have a hard time explaining what and why they did something. And that explanation becomes more difficult the further (distance and time) you get away from the actual behavior. I have found that interviewing customers can help gain a better understanding of consumer perceptions, which can help improve a process or policy to improve the overall customer experience.

    There is a great book called consumer.ology by Philip Graves that provides some great examples (e.g. Red Bull, and Coca Cola).

  • Thanks Scott and I agree. I think a “side effect” of this activity is that it sends a message to customers that you care.

  • Cindy C.

    In person meetings help with not only what you mentioned but also the hidden factor of nonverbal communication. Just started reading Louder Than Words by Joe Navarro last night. That may be influencing my thinking right now.

  • Really great point Cindy. Thanks for sharing that!

  • Gerry Michaels

    Mark, you had me at “go out and talk to them”, what a concept. So many forget that there is nothing like that irl contact. Just last week I drove an hour and a half to meet with a client for an hour. They had moved into a new office and gotten some great new technology tools for their practice. It was a very productive hour.

    I am a strong advocate of irl meetings whenever possible, and on a regular basis. There just is no substitute, yes technology tools are helpful, but not the end all of interaction. Nice, I liked this one a lot. Thanks

  • I totally agree that word clouds are no replacement for conducting research the traditional way–have real conversations with real stakeholders. But used in tandam with human analysis they can add another dimension your findings. Besides, they look cool.

  • Hmm, we actually have word clouds as part of Pulse Analytics reporting. I always show it because people like to see it…and I have seen it identify major cases of “you don’t know what you don’t know” – where people are talking about aspects of your brand that you were not explicitly looking for or measuring.

    Is sentiment analysis the “end all” absolutely not…and I can see the lure of becoming “lazy marketers” by letting the monitoring product do all the work without doing further manual digging. However, it is still good at broad brush analysis of major topics/issues associated with your brand. Good tools also support real time monitoring as well as alerting features to immediately alert your decision makers if keywords like “boycott” and “dangerous” come up.

  • Many thanks for the very nice comment Gerry!

  • There are exceptions to every rule and you are smart enough and experienced enough to know when to use the right tools for the right problem. But believe me when I tell you most of the time they are abused as a substitute for something that will give us a real direction! Thanks my friend.

  • “Abused as a substitute…” – good point. No substitute for hard work, right – unless you substitute EFFICIENT hard work!

    Thanks, Mark.

  • Pamela. J. Watkins

    I agree with Mark. It’s time we get back to the basics. Great article.

  • Pamela. J. Watkins

    That’s actually one of the issues Mark, the new generation marketers. They are being taught that costly research, useless time conusming reports and word clouds are the NEW and improved way. WE CAN do our best to gently nudge them in the right direction

  • kremsa

    Another set of tools that are popping up for today’s marketer are automation tools. We use starsocial to automaticaly send customers simple surveys after they took specific actions on our site or didnt.

    For example: one of our apps lets businesses create polls. Its free yet we were getting a lot of drop off so we now send emails to users that do not complete the creation within 15 of intending to create a poll.

    The response is great, its all over the board but a few experience the same issues, very specific problems that we would never think about like the lack of understanding of what the poll will look like when its done or inability to figure out a feature.

    This allows us to improve what matters.

  • Pingback: The best of @living_business for 7/30/2013 - Living Business()

  • correlationist

    Simple, yet profound – that typifies you, Mark. As someone who has sold social media data & research (as opposed to pure social media monitoring plays) for the last 3 years, I can say these tools can be useful, but it has to be applied to specific objectives. Visualization aside, there is no denying the power of social media monitoring and research to drive real value across the entire business ecosystem. In fact, social media can, and is, a great source for deep category and early stage exploratory insights. The real issue is that legacy methodologies continue to dominate and very few folks have successfully looked at social media data through the rigors of a research lens.


  • Hey Mark, I’m going to have to agree and disagree with you at the same time.
    I want to agree with you that you can’t always rely on just these automated tools for all the answers you’re looking for. I work for one of the tools you mentioned above (Marketwired’s Sysomos), but I always tell people that while tools are helpful, they can’t do all the work for you. Marketers (or anyone else using our tools) still needs to put time and effort into digging deep into what they find in a tool. And that may come from more work within the tool or outside it. But, I do agree that people need to put in the work.
    That said though, tools that show things like our word cloud or automated sentiment ranking can still be a great place to get you started on that work. Sentiment analysis is probably never going to be 100% accurate (there’s just too many nuances in how people talk/write), but it can be great for getting a high level overview of where your brand stands (or whatever it is you happen to be looking at). The same holds true for word clouds. Word clouds will not give you everything that your audience/customers are thinking, just what they are writing. However, by seeing what issues they write about the most can give you a great idea of where to start and what to start looking at.
    So, I do think that these tools can be very helpful, but you can’t rely just on the tool. People do need to put in the work and get to the route of what’s happening and what people are thinking. These tools that we make are there to help you along the way though.

    Sheldon, community manager for Marketwired

  • Pingback: Another Lesson in Church Math | 2x2 The Church Without a Building()

  • A major problem with social media data is they come from… social media. Only a minority of customers express themselves about brands online and those doing so may have specific psychological traits. Usefull insights can be found with social media research, yet we could miss a bigger picture by relying on the ´sentiment´ of a specific category and ignoring the vast silent majority.

  • Pingback: SEO content marketing roundup, week ending July 31st()

  • Pingback: HWPCC News » Blog Archive » SEO content marketing roundup, week ending July 31st()

  • rhonda hurwitz

    This is a topic close to my heart. I would argue that the MOST important activity for a marketer is gaining a deep understanding of the customer. In my experience, social media gives us a great opportunity to listen, at a granular level, if we take advantage. The comments and conversations you can find in forums, and online product reviews, might yield that voice of “one”, but you have to comb thru a lot of nothing to get to it:)

  • Sounds like a plan Pamela!

  • Hey Prince. Good to hear from you. You point out a real irony in the market today. We have SO MUCH DATA! It is an incredible opportunity yet we often overlook what’s there for the immediacy and comfort of a word cloud. Crazy!

  • Hey Sheldon, first I want to be clear that I wasn’t knocking your product at all. In fact, I just did a project with MarketWired. I also recognize that wordclouds have a (limited) place in the world of marketing research if all you need is a quick snapshot in a limited time period — although I would contend that if you really want that, there are better ways to display the data, like a Pareto chart. I think the opportunity for your company is to help teach marketers what they can really do with the data you provide. Is it meaningful? Is it relevant? Is it statistically-valid? Are there confounding variables they should be aware of? Asking those questions — and getting the right answers can help lead to real insight and competitive advantage!

  • Very true! We must be aware of context and also that a few aggressive types can hijack the conversation!

  • Boom. Good summary Rhonda!

  • I totally agree (and I didn’t take it as a knock).
    That’s why something like a word cloud isn’t the only text analytics tool in our belt. Things like our buzzgraph (which shows how words are actually appearing together) or our entities tool (which is able to pick people, places and things driving conversations out from all the chatter) are much more useful. However, everyone still seems to think that the word cloud “looks pretty” for their reports.
    As well, I know that we try to work with all of our clients to make sure that they’re getting the most out of our tool for what they’re trying to achieve. (I hope you had that experience)

    I really just wanted to pipe in and say that they’re not totally worthless.

  • Christine Webber

    What you highlight really well here is the need never to forget the importance of face to face communication. The tendency these days is to rely far to much on technology. This was highlighted for me at the Oi Social Media Conference in Cardiff, a few months ago, where you were the main speaker. Well over half the audience were so engaged on their mobile devices that they were paying little or no attention to the people around them! What a wasted opportunity to develop new relationships.

  • Pingback: Cooperating with the Competition – Inkling Media()

  • Pingback: Which of Your Customers is Most Important? or, What You Don't …()

  • Pingback: Customers Don’t Think Like Marketers()

The Marketing Companion Podcast

Why not tune into the world’s most entertaining marketing podcast that I co-host with Tom Webster.

View details

Let's plot a strategy together

Want to solve big marketing problems for a little bit of money? Sign up for an hour of Mark’s time and put your business on the fast-track.

View details