Evaluate Weigh the pros and cons of technologies, products and projects you are considering.

Content analysis tools earn chops in research on social media data

Many organizations are discovering that content analysis tools can help validate market research efforts and hone their messages by mining customer-sentiment data available on social networks.

It may be an understatement to say that content analysis tools are being used in a variety of ways by a growing cadre of users. The set of technologies that makes up the content analytics universe amounted to an $835 million market for software and support services last year, according to a new report by consulting firm Alta Plana Corp., which said it expects the market to increase at an average annual rate of 25% for the next several years.

The user spectrum includes municipal police agencies that use text analysis software to analyze databases and written reports for crime-fighting purposes. But in large part, the market growth is being spurred on by companies like Hilton Inc., which is using text analytics to collect and act on suggestions from hotel guests, and a North American bank that uses content analytics software to develop new services to offer customers. (The bank asked not to be named.)

While enterprise search technology looks for content you already know you have, content and text analytics tools look for useful information that often is hidden in documents, emails, digital files and numerous other types of unstructured content. And that includes content repositories on both sides of the corporate firewall: Many organizations are discovering that they can both kick-start and validate their market research efforts and hone their marketing messages by mining and analyzing the vast treasure troves of customer-sentiment data available on social networks.

For example, J.D. Power and Associates, a consumer and market research outfit in Westlake Village, Calif., is using text analytics to examine market possibilities for its clients, which include automotive manufacturers, electronics companies, insurance and financial services firms, and makers of consumer goods. One of the latter, PepsiCo Inc., recently hired J.D. Power to study the non-alcoholic beverage tastes of what has been dubbed the Millennial Generation, a group roughly defined as being between the ages of 17 and 29.

After analyzing social media data, J.D. Power found that after coffee, juice was the top choice for people in that group. And by using text analytics software, the company collected in three months what normally would take a year or more to obtain via a traditional study using focus groups and other conventional research methods, according to David Howlett, head of consumer insights and strategy at J.D. Power.

“One of the benefits of [text analytics] is that speed of research is becoming more and more a key issue,” he said. Because of the results of the study and the speed with which it was completed, Tropicana, one of PepsiCo’s brands, is now placing juice at convenience stores in an attempt to put the products in the path of on-the-go millennials. “It’s informed a tangible marketing strategy that should result in increased sales,” said Howlett.

Marketers across the consumer goods industry are keeping track of what people are saying on social networks and in customer service calls and emails, using natural language search and text analysis technology to look “for patterns and relationships in those channels,” he added. “It’s good for exploration and discovery and helps uncover new needles in the haystack.”

In another example of how natural language processing has been used for consumer research, J.D. Power studied the use of the word “burn” in social media. It started with the assumption that the word had negative connotations, but Howlett said the company soon discovered that with regard to Listerine mouthwash, customers were using “burn” in a positive way.

Funneling feedback via content analysis tools

This form of research has additional value over traditional forms because it lets companies gather and analyze “unaided and unprompted consumer feedback,” Howlett said. “By looking at the type of language they’re using, you can start to understand what problems and needs these people have.”

Another aspect that makes content analytics so powerful for consumer research, according to Howlett, is that it allows for more effective surveys. Once the bane of marketers designing surveys, the open-ended question now offers new benefits to researchers because of the ability to use content analysis tools to analyze and categorize written responses. Howlett said that often results in shorter surveys with higher-quality results make it possible to do both quantitative and qualitative analysis at the same time.

And in some cases, the new tools are taking the place of other research methods altogether. Organizations such as Procter & Gamble and Unilever “have already begun to phase out focus groups and replace them with content and text analytics,” Howlett said. Companies are also seeing potential savings in using text analysis software to replace expensive telephone and mail-based surveys, he said. 

“I think it’s a really flexible technology, so it can be used for market research, certainly,” said Leslie Owens, a content management and collaboration analyst at Forrester Research Inc. in Cambridge, Mass. “It’s just about coming up with the right way to use the technology to discover all these different voices – to do qualitative research to get a sense of what people are saying, what their passions are.”

In addition, applying content analytics technology to social media data enables public relations departments to respond quickly to issues surrounding a company’s name, products or business practices, according to Owens. In such cases, businesses can use text analytics tools to “listen for early warnings of a problem,” she said.

Dig Deeper on Text analytics and natural language processing software

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.