Posted
May 2009
People are Talking
Anticipate product performance through online discussions
Based on the Research of Shyam Gopinath, Jacquelyn Thomas And Lakshman Krishnamurthi
Talk may be cheap, but listening to what people are saying about your product can be a valuable method of improving corporate performance. According to recent research, there is a measurable connection between what is being said about a product in online posts and real-time customer behavior. Becoming aware of online comments and learning how to use the information can avert potential downturns in sales and can help companies fine-tune their marketing.
Potential buyers of a product are influenced by existing users of that product, the research shows. According to Lakshman Krishnamurthi (A. Montgomery Ward Professor of Marketing at the Kellogg School), “If you ask people what’s most important to them in evaluating a product, they say, ‘What other people like me say about it.’” Krishnamurthi and his colleagues, Shyam Gopinath (a doctoral student in marketing at the Kellogg School) and Jacquelyn Thomas (Associate Professor of Marketing at Southern Methodist University), examined online word-of-mouth for the cellular phone industry. They focused on five specific brand models from five leading cell phone companies in the United States. Using data from an online forum with more than eight million posts, they explored the conversations of individual posters over time and analyzed how the nature of these posts related to individual customer behavior. From there they examined how the nature of online conversations relates to corporate performance.
Classifying Attitudes
In a working paper based on their research, Krishnamurthi and his colleagues report that they developed the data set by identifying keywords in the posts that expressed an attitude toward a cell phone and usage experience. As Krishnamurthi explains, “We classified people’s comments in these posts in three ways. One is an action-type statement, such as ‘I’m going to buy it.’ Another type expresses emotion, such as ‘I hate it.’ The third category is made up of attribute-type statements that have to do with quality, things that relate to the functionality of the product, such as ‘It has great reception.’” Each type of rating can be positive or negative.
Becoming aware of online comments and learning how to use the information can avert potential downturns in sales and can help companies fine-tune their marketing.
Using specially designed software, the researchers rated the action, emotion, and attribute statements on a scale. According to Krishnamurthi, “It’s a little bit like artificial intelligence. You take a large number of posters and look at all the words they use, and create a classification of these words as highly negative through highly positive. The software has a dictionary, and when these posts are made the software automatically classifies them on this continuum.”
Buzz Action Score
With their data the researchers developed a customer-level metric based on active and passive customer behaviors, which they dubbed the “buzz action score.” Using the buzz action score from individual posters, they derived two aggregate-level metrics: the “buzz index” and the “buzz share.” Evaluating the buzz index and the buzz share allowed them to demonstrate how online word-of-mouth relates to actual market performance. Figure 1 illustrates the conceptual framework for the process. First, an existing user posts an evaluation. Next, a potential purchaser encounters the evaluation. As a result of the online “conversation,” the potential purchaser either does or does not buy the product. The results of the “buy” or “don’t buy” decisions resulting from the conversations ultimately influence the company’s performance by showing up in sales data.
Figure 1: Online word of mouth - conceptual framework
Previous research has shown that a relatively small group of people can have a substantial influence on the majority. The researchers found that the same is true in online communities. The highly influential posters, whom Krishnamurthi and his colleagues call “mavens,” tend to have an especially intense interest in the brand, either positive or negative. Because of this intense interest, they are more likely to advocate a strong opinion, which in turn has a much bigger influence on potential purchasers than does the opinion of the average user.
Brand Performance
So how does that influence relate to measuring brand performance? According to Krisnamurthi, “The traditional way is to look at sales and look at marketing and try to see if marketing has been effective in moving the needle.” However, a problem with actual sales measures is that they are delayed. For example, phone manufacturers only know what they are shipping out to vendors. They will not know actual sales figures until some later point in time.
In contrast, by monitoring online postings, a firm can find out much sooner if there is negative sentiment among customers about the brand. The researchers found that the buzz index, whether positive or negative, is a leading indicator of sales. Krishnamurthi explains, “There is very practical value to tracking these online posts. You don’t have sales data for three months from now, but you have the online postings right now. This makes it possible for companies to fine-tune their marketing and not wait for the sales to go down.”
The researchers also found that the importance of the attribute and emotion scores varies among brands. “Let’s say you’re a communications company, and your marketing has been pushing the coolness factor,” Krishnamurthi says. “Then this is a gut check to see if your marketing is consistent with how your customers are reacting to it.” He added that the finding about variance of attribute and emotion scores among brands has far-reaching implications. For example, “Say I am brand A, and you come to my Web site. One of the questions I’m going to ask you is, ‘What brand do you own now: A, B, or C?’ If you say ‘B,’ I know what I’ve found in terms of whether attributes or emotions are important to brand B customers, and I’m going to lead you to a portion of my Web site that talks to you about my product in terms you might like.”
The authors suggest that firms should monitor online word of mouth in order to “stay ahead of the race.” They add, “An Internet-armed consumer can become your greatest asset or your worst nightmare.”
Read more articles in:
Entrepreneurship & Innovation
Marketing
Print without comments
Print with comments
15 Comments
May 5 2009
I agree! I have been working with www.chatmine.com for the past several years to help various clients (B-B and B-C) be a “fly on the (virtual) world” to mine online Consumer Generated Media (CGM). This research tool is indispensable!
May 5 2009
Nice work and nicely written. I’m a Kellogg ‘06 grad and I run a social media company. We are seeing an increased interest from our clients in “listening” to the social media sphere. Companies large and small are beginning to realize what your research has proven.
May 8 2009
A nice analysis of an emerging trend. Interesting to add that I read also where IBM research has developed a very sophisticated and realtime tool called COBRA to analyse these Web 2.0 data sources in many different native languages. The article said that some of the largest firms globally like Consumer Package Goods, Banks and Telco’s are using it to improve their response to customer satisfaction, improve product performance and shape future marketing campaigns based on what people are interested in and talking about. This looks like just the beginning of a huge trend built on the back of social networking.
May 11 2009
While I have seen increasing interest in participating in and driving online conversations by companies, the measurement has lagged behind. We are already on to the second generation of these tools, such as Wise Window, but many marketers are reluctant to spend money to get substantial data, perhaps because they are unconvinced about its value.
May 13 2009
Very interesting post. Wanted to introduce you to Social Radar, a social research application that can be used to analyze and measure web chatter from millions of sources (blogs, Twitter, social networks, news feeds, etc.). Social Radar enables users to:
-Trend conversations and buzz to measure the reach and effectiveness of campaigns
-Gauge positivity and negativity
-View keywords and topic clouds to understand what’s being said by millions instantly
-Locate key influencers
-Analyze over 2 years of historical data
For more info, visit www.infegy.com/socialradar.php or check out our blog www.buzzstudy.com.
May 13 2009
Nice to see this research come out - we from MotiveQuest LLC (in Evanston) were happy to support this research with the underlying data, software and analytic support.
Great stuff - we have more on the link between advocacy (what people say to each other) and sales here:
http://tinyurl.com/6oc4rl
Tom O’Brien
CMO
MotiveQuest LLC
tobrien@motivequest.com
May 15 2009
It is great to see some quantification of that old say “word of mouth is the best form of advertising.”
May 15 2009
With evolving standards of Web 2.0 these things will gain more importance. In web world there are efforts to consolidate all such data in systematic way. we never know it may become tool used by almost everyone 50 years down the line and there may be a popularity index for everything that will go up and down everyday :) Web is still way to go !!
May 16 2009
This is the second research initiative I’ve read about that suggests online buzz and activity can be a leading indicator of future sales.
I think business leaders, including myself, will be looking for more case studies that prove there is a direct link between online conversation and sales.
Your readers may be interested in my ongoing book report on Groundswell. There are real life case studies on how social media has/is impacting businesses.
In fact, chapter five talks about social media chatter and buzz being a leading indicator of sales of Mini cars in the U.S. MotiveQuest was the backend app/service used for this I believe.
http://marketingthatmatters.blogspot.com/search/label/groundswell
I have no connection to MotiveQuest, Mini or Forrester.
Jun 15 2009
I am a college student in China.I cannot understand some parts of this article,like the “buzz index” and “buzz share”.Can some one help me with these?Thank you.
Jun 15 2009
I’m not sure that I get it. Why would marketers—in an interactive age—settle for secondary feedback? No offense intended but I find the entire notion of harvesting/mining the conversation database to be rather backward—mass marketing influenced.
Why are we even bothering with this approach? This approach is ONE WAY—does not leverage the inherent interactivity of the Web. Wouldn’t it make more sense to simply ASK customers in real time and study the DIRECT feedback??
Thanks for considering.
Jun 20 2009
@Jeff: I agree that asking customers would create valuable answers as well. And you can direct discussions to what you want to know, by asking the right questions. Something that you can do by analyzing User Generated Content (UGC).
But I think both approaches are valid. It is not one or the other, it must be both. But to ignore what is said online about your brand, your products, etc. would be stupid. The information is there, it just has to be found. It might not be representative, but it can harm or help your company. It was written by one, but maybe read by thousands. This is an accelareted word of mouth as you know it form before. While you my have talked about your experience with a brand with 10-15 people you know, now it might be read by 100.000, and it is there for basically forever. While you might calm down, your rant about a problem might stay for years on the internet.
I think it surely is worth considering to have a look at online buzz. It might not be representative, but it can give you a good idea what is going on out there. As a big company, I would at least install some kind of “alert system”.
Jun 22 2009
Who Cares?
Literally…‘who’ are the people who ‘care’ about my product? Why do they care? What need is it filling and is it delivering on that promise to a particular meme/target audience?
These are the questions that are important to me as a consumer and as a marketer.
Direct Marketers have known for years how to use testimonials and convincing copy to make sales. Targeting, regression analysis, database marketing are known knowns at this point. So the technology and landscape evolves and shifts but the result is often the same - unconscious consumption of useless stuff - or not.
The question (to me) today is - how can consumers and businesses differentiate between all the available offerings and means of ‘‘pushing’ product through a pipe to create increasingly higher levels of value for the individuals, groups and societies that need (not just desire) them?
Jun 23 2009
@caoyuan:
Buzz Index reflects the online word of mouth conversations of a specific brand, whereas Buzz Share for a brand is a relative measure which takes into account the nature of online conversations of the competing brands as well.
Jul 6 2009
OWOM—-> this is one of the interesting topic..great work…WOM had been very important in Asia countries’s marketing strategy. It the first source of information that consumers trusted.
Online??? will need to wait and see if any reliable information can come out from this idea.
Krisakorn Sukavejworakit
Bangkok, Thailand