March 21, 2017
Dirt Devil has seven categories of products. Each of those categories has up to 28 different products within it. So it’s not enough to say, “You should buy a Dirt Devil.” That will only confuse the customer when they go to the website or march up and down the vacuum aisle looking for the recommendation. If there are dozens to choose from, how is the customer to know what’s best for her (or him)?
As Malcolm Gladwell so eloquently explained in a now famous TED talk, at one point in time, customers wanted choice. Diversifying it’s product offerings made Ragu a mint. Then there’s more recent examples about the paralysis of choice that indicates too many and customers will not buy anything.
We looked at some conversation research around the Dirt Devil brand and discovered that while we cannot pinpoint a correlation between too many choices and hard sales data in online conversations, we can observe how Dirt Devil customers refer to the products they use.
While the categories led the way — indicating consumers are more apt to describe their Dirt Devil in broad form — there were several attempts at identifying specific product names and models. AccuCharge and SimpliStick were among the top six results of product identifiers in our research. That could speak to strong branding for those products. But why don’t others emerge? Does the brand divide its marketing among different agencies or marketing initiatives? Is that why some standout and others don’t?
Online conversations my not tell us the answers to those very specific questions, but a hearty conversation internally might.
And is it more beneficial to have everyone referring to Dirt Devil stick vacuums as such rather than in hodgepodge ways of reference? One hundred people shouting praise on a Dirt Devil stick vacuum is probably more beneficial than 24 complimenting the SimpliStick while 15 talk about the Power Stick and nine refer to the Power Air Stick, right?
This exercise is not to imply that Dirt Devil has a branding problem or unnecessary confusion among consumers about what products they offer. It is simply a way to open a dialog about the Paradox of Choice and whether or not branding initiatives could or should solve for it.
Word-of-Mouth Marketing is up to 200 times more effective than advertising, according to the Word-of-Mouth Marketing Association. Shouldn’t your brand’s conversation focus then be on unifying how people talk about you so you can deliver a more consistent wave of conversation when they do?
It’s certainly good food for thought and something you’ll never get a grip on unless you’re studying the online conversation about your brand. If you need help doing that, drop us a line. We’d love to help.
March 14, 2017
One of the most common forms of data visualization among social listening softwares is the tag cloud. The graphic representation of which topics are the most common organized by word, size and color is easy for the layman to decipher so it is dangled at the end of the software company’s string like top sirloin.
But it’s just a chicken nugget. Or, more aptly, just the breading around the chicken nugget.
Topic wheels are a bit more informative methods of data visualization. They enable you to see subtopics easily. But the data visualization is still just a superficial layer around the insights your data contains. They help you see one layer down.
But insights are seldom found one layer down. Understanding of the conversation — the why behind the emerging topics and themes — means drilling down deeper.
Take our study of Dirt Devil, for instance. We may notice looking at Tag Clouds and Topic Wheels of data visualization that durability is an issue that surfaces in the negative conversations around the brand. But why does Dirt Devil have durability issues? To know that, you have to drill down into the negative, then into the durability topic, then analyze and understand the various issues there.
The level of detail that can provide a product manager with actual insights to improve the product is not found using a tag cloud or a topic wheel. It’s found by diving in and analyzing and understanding the full context of the conversation. With this information — certainly represented visually for ease of understanding — we can tell the product manager that there are structural issues in quality of construction, weakness in the unit handles and motor issues, particularly when used for pet hair. These insights give the product team direction so they can either A) Ask deeper questions in further research or B) Focus on the opportunities to improve the product.
The overarching point is that if you’re relying on visualizations of your data rather than analysis of it, you’re missing a lot. In fact, we would surmise you’re missing everything.
We would love to help understand your data. Want to know more about what customers say about your brand? Your products? What you can do better? Drop us a line. We can help.
March 7, 2017
Chief Marketing Officers (CMOs) are spending more on analytics now than ever before, but also admit that barely 1/3 of the data they’re paying for gets used. That’s according to The CMO Survey from the American Marketing Association, Deloitte and Duke University’s School of Business. One of the biggest reasons CMOs aren’t using the data? They say it’s too complex, lacks insight and relevance.
This is exactly why we started the Conversation Research Institute. No, we’re not going to solve that problem for all aspects of marketing analytics. But when a CMO gets a report from a social listening platform, it’s a vague assortment of charts and graphs. It doesn’t explain the WHY any of those bars are as big or small as they are, the pie chart looks the way it does or the colors are one way or another.
What Factors Prevent Your Company From Using More Marketing Analytics?
When you pay for software as a service, all you typically get is the software. The service part of it doesn’t refer to someone to serve you insights or make the software work for you.
CRI is focused on taking either your existing social listening software or implementing the software we use on your behalf, then interpreting that data so it is:
- Easy to understand
- Delivers insights you can use
- Focuses on the voice of your consumer to deliver relevance
CMOs are too busy and have too little time to interpret the data they receive. A strong analyst is going to see that and deliver what the CMO needs when he or she needs it. They’re going to focus on the stakeholders in question and on the issue of relevance. Without those two focal points, no amount of data or charts or graphs will help the CMO make decisions.
It is true. CMOs are spending more money on analytics. According to the study, analytics will jump from around five percent of marketing budgets to almost 22 percent in the next three years. Why on earth would they pay more money for something they use less than 1/3 of?
We owe it to ourselves as analysts and evangelists for conversation research, social listening and social analytics to close that gap and ensure that CMOs are getting their money’s worth. We know what we’re doing about it at CRI. What are you?
February 21, 2017
For those interested in the world of Artificial Intelligence and Conversation Research (which is driven by A.I. algorithms), IBM Amplify 2017 is a must-attend event. And I’m pleased to report that I have been invited to speak as part of the event’s innovation leaders series.
The event will be March 20-22 at the MGM Grand in Las Vegas. You can register online at http://ibm.com/amplify
My talk, which is slated for March 22 at 10:15 a.m. local time, will focus on the need for human analysis to close the gap on Artificial Intelligence and its usefulness when used to make sense of unstructured data. As with most talks I give, it will raise a few eyebrows, but hopefully push the industry forward in building A.I. that works better.
And for those of you interested, I can score you a VIP invite to an influencer dinner with myself, Jay Baer and others. Just drop me a line before you register and I’ll tell you how to score that invite!
CRI is excited to be represented at what is essentially the thought leadership home for A.I. and it’s conversation research offshoot. IBM Amplify is essentially the user conference for IBM’s fabled Watson A.I. engine. To be included is a nice honor for both CRI and me.
See you in Vegas!
February 16, 2017
The fun for me in analyzing online conversations is the proof points the data provides. No longer do product, experience or marketing communications decisions have to be left to assumptions. The data allows you to turn them into assertions.
In our recent report on senior living, we analyzed online conversations of people discussing the major types of senior care facilities. We found hundreds of conversations mentioning nursing homes, assisted living facilities, independent living facilities and long-term care options. We broke each of those conversations down by facility, sentiment and topic.
When you do this, you get a glimpse into what consumers truly think. Not only are we not prompting them for answers, which in and of itself biases the information, but we’re simply recording when they talk about the topic in question voluntarily and freely.
What does this type of analysis tell us? Take for instance this visualization:
This is a breakdown of the conversation topics within the posts we categorized as focusing on assisted living facilities where the main topic was the experience of the family of the patient (which is important since the primary buyer is the adult children of the patient), and those experiences were scored as having a negative sentiment. So 30% of all negative conversations about assisted living facilities (represented in the circle to the left) were determined to be about the family experience. The right hand circle breaks those down by specific topic.
What this tell us is that 32% of the negative family experience conversations were about shopping for the facility overall. What is it that is so bad about it? We’d need to move a layer farther in analysis to discover that, but since we have the data, we can! Another 32% mentions they prefer an alternative to an assisted living facility. Further analysis shows that they don’t prefer independent living or nursing homes, but rather staying home and not needing a care facility at all.
While this may seem a logical conclusion if you understand the consumer, that has not been statistically proven before, to our knowledge. Now it has. But that insight can also give assisted living marketers more pointed insights to develop better copy, sales materials or even sales strategies, enhancing conversions and driving more customers.
Emotions while enrolling and family in-fighting are significant portions of the negative family experience, too. What can that tell an assisted living marketer hoping to land more clients? Those conversations can be further vetted to see if common threads run throughout.
The more you peel back the layers on analyzing online conversations, the more interesting nuggets you discover to fuel decisions for marketing, user experience or even product development. And those can build a better, more profitable brand.
The only question left to answer is why haven’t you started?
For more analysis of online conversations around the senior living industry, including a mapping of the buyer journey for senior care, see our Conversation Report. For more about how CRI can help you in analyzing online conversations around your brand or market, drop us a line.
January 5, 2017
It’s one thing to know what percentage of a given audience is male-female, different ages, ethnicities and so on. It’s another to understand how that audience compares to the norm. Indexing a given set of results against a generally understood or accepted point of reference not only frames the context of that audience characteristic, but can help you elevate important insights in conversation research.
Some social listening platforms offer audience indexing in the demographic and psychographic data. This seldom used and often misunderstood statistic is one we constantly refer to at CRI since it can lead to more intimate understanding of the overall make up of a given audience.
To better understand indexing, take a look at this chart on a given audience’s ethnicity. Its primary function is to show the percentage of the audience broken down by ethnicity.
But we’ve also displayed the index compared to the general demographic profile of a commonly used site (in this case, Twitter). We know from multiple resources (Pew, Northeastern University, etc.), in general, Twitter’s audience parallel’s the U.S. population in terms of ethnicity. Even with some variations considered, at a minimum, we are comparing our audience to an audience of people who are active social media users.
As you can see in this audience, caucasians index at a 1.14 rate. That means that this audience if 14% more likely to be caucasian than the base audience of Twitter users. So it skews white. It is comprised of slightly more African-Americans, 19% less Asian, a bit more less American Indian or Native Islander and “other.”
But look at the Hispanic index. An index of 0.28 means this audience is almost 80 percent less likely to feature Hispanics than the base audience of Twitter users.
What does this tell us? It could tell us a few things:
- Hispanics aren’t talking about this topic (if you’re doing conversation research) or buying this product (if you’re analyzing sales data)
- The industry or brand in question does not appeal to Hispanics
- The industry or brand in question ignores Hispanics
The definitive answer would require more detailed research, but seeing the huge disparity in the indexes gives us reason to investigate and perhaps an opportunity to fuel decisions to improve the business.
And keep in mind that demographics aren’t the only thing that can be compared in index form to Twitter or other data sets. You simply need a known and common data points. In CRI’s research, we frequently surface indexing for age, gender, ethnicity and geography, but also social interests, professions, bio terms and more.
Indexing is a powerful statistical feature to understand as a researcher or a marketer. Understanding it could be the key to unlocking equally as powerful insights for your business.
For help with understanding your audience and how they index compared to known audiences, drop us a line. We’d love to help.