How to Analyze Consumer Generated Media

Any analysis of what is being said by consumers across electronic media (a.k.a. Consumer Generated Media, or CGM) can lead to spurious conclusions if one doesn’t understand the data. For instance, how is one to know if a given post is representative of wider consumer opinion, or just an outlier? How does one know if a given post is going to be read by others, impact their perception, or influence a purchasing decision?

In order to answer these questions, one needs to understand what CGM is, who is generating it, and how it affects consumers. This article will attempt to guide the reader through the first steps of the process.

The types of CGM

Broadly, CGM sites fall into two categories: news/information/opinion and subject/issue. Significant overlap between the two exists, but for illustrative purposes, two high-level categories will suffice.

News/information/opinion sites (aka meta-sites -- which gather links and stories together from other sites) tend to be either collaborative pages that post links from other websites to a forum, often including a running commentary/opinion component, or a sole author’s page that posts links from other websites and contains his or her own opinions. As a general rule, such sites tend to be weblogs. While some news sites do create original articles, the majority tend not to generate many stories, but rather rely upon commentary about stories from other sources. As with traditional media, this pattern tends to create something of an echo-chamber effect, with a post or link rapidly spreading among sites. However, these sites should not be thought of as simply hyping stories from traditional media. While the content might not be original, many sites source stories or links from each other, enabling a post on a smaller or obscure site to rapidly become an active and visible issue.

Subject/issue sites, (under which banner I’d include meta subject/issue sites such as Epinions) tend to focus on products, product categories or issues. The users of these sites tend to be passionate about a given topic, and want to express their opinions. These sites tend to have long running discussions, and are often where consumers turn when they’re seeking advice or information. As a general rule, these sites tend to be message boards (or Usenet groups).

How to prepare for a search

Now that one knows the types of sites, we can talk about how to find them. There are two means to analyze and keep track of CGM: the traditional manual method; and a more sophisticated automated system. Which one is more appropriate will depend on what one is trying to accomplish, the volume of posts about a product, company or issue, and the scope one wishes to examine. The broader the goals, the larger the volume, and the greater the scope, the more one needs an automated system to distill meaning from the content.

To determine which approach is right, one should start at a search engine, such as Google or Yahoo!. Enter in keywords that you think would occur in a post that addresses an issue about which you’re interested. This could be a product, a company or an issue. Chances are good that within the top 20 results returned, there will be some CGM content, and deeper into the search results there will be more. (We’ll get into the discussion of site importance later, but it’s worth pointing out here that these first sites are very important. They’re what consumers who look up a topic are going to find too, and may even be an individual’s first impression of a product.)

If one wishes to focus solely on CGM, there are a number of “alternative” search engines, including Technorati.com, Waypath.com and Daypop.com, which look at weblogs, or a service like BoardReader.com, which searches message boards. The Usenet can be searched trough groups.google.com.

If there are only a few posts at the above-mentioned sites, and one isn’t interested in examining competitors or marketplace trends, then it is likely most of the analysis can be done manually. However, if there is a large volume of discussion, an automated solution may be necessary to help users sort, categorize and mine posts for information. A few services, including Cymfony’s Digital Consumer Insight, provide systems for automated analysis of CGM content. These applications use continuous search feeds to pull in content from all types of CGM content, and the deeper services apply a secondary level of analysis to the postings. This generates meta-data about the content, allowing a user to filter out noise and spam quickly, spot trends and differentiate them from aberrations, ensure statistical validity of one’s analysis, pull out common themes, issues and discussion tone, and segment postings for more effective market research.

Determining a post’s importance

As with the first wave of digital individual content (personal web pages), the same rules of information overload still apply. Limited by time, attention span and interest, people still tend to congregate at certain primary locales, even if those locales are even further fragmented than they were a few years ago. There are three characteristics that make a site important. The first is the size of the audience, i.e. number of page views or impressions. The second is a correlated statistic, inbound links, i.e. the number of other sites which link to a given site. The third is the audience demographics.

A handful of weblogs command a disproportionate share of all CGM viewership. The weblog audience is not evenly spread across all blogs and boards, nor does it even follow a linear distribution. Page views follow a hyperbolic distribution, with the top sites garnering a massive share of viewership. However, one should note that viewership can be highly fluid. Although some sites or forums have consistently high numbers of monthly visitors, the viral and interconnected nature of the web can cause a previously infrequently visited site to experience exponential growth. This fluidity is best understood by examining inbound links, which are a good indication of traffic to a site at any given time.

With regard to audience demographics -- although a site may be unpopular with the general public, it may be highly influential among a specific segment. For instance, a discussion board site may only have a few thousand viewers a day, but may be the top Google ranked site for photo printing enthusiasts. While this site might not be important to the population at large, it’s highly important to a digital camera manufacturer, or any consumer in the market for a digital camera. Given the large percentage of people that research products online, the opinion of just a few people on a key site can have a significant impact on purchasing decisions or corporate reputation. A post on a top site is guaranteed to get a lot of viewers and have a significant impact on corporate reputation, but a post on a smaller issue-driven site may be more important for consumers.

As a general guideline, any post appearing on a top 10 weblog could be called influential. Cymfony’s current top 10 CGM sites are:

  • Boingboing.net
  • Slashdot.org
  • Plastic.com
  • Metafilter.com
  • Dailykos.com
  • Instapundit.com
  • DrudgeReport.com
  • AndrewSullivan.com
  • TalkingPointsMemo.com
  • Gizmodo.com

Reading these sites will allow one to keep on top of burgeoning issues. However, if one wants to get ahead of an issue or story, and prepare in advance, one needs to continuously analyze as wide a selection of CGM sites as possible.

Every industry or interest will probably have its top lists too, something we wish for our corporate blog: http://www.cymfony.blogs.com/

Reading and analysis

What one should be looking for within any given post depends on what one is trying to accomplish. The information useful to someone concerned with corporate reputation is different than that of someone in competitive intelligence, which is different still from that of someone in marketing or engineering. Although there isn’t enough space to provide that information here, I can provide some tips to keep in mind to get the most out of your CGM analyses:

  • Don’t go in with preconceived notions.
  • Listen to people’s responses. Why are they upset? Why are they happy?
  • Read the context to understand why people feel as they do.
  • Follow links if provided.
  • Try to see things from the author’s point of view.
  • Become involved in the process -- you can even ask for clarification and guidance.

Most importantly, if one is struck by the need to engage with an author, be very careful. Assume anything you write or express will be published, and follow Jeff’s cardinal rules for CGM: Be Open. Be Transparent. Be Honest.

Jeffrey Feldman is manager of consulting and strategic services for Cymfony. Andrew Bernstein, CEO of Cymfony, contributed to this article.

 

Comments