Social Media & Surveys: Mining Sentiment & Attitudes to Understand the New Breed of Customer

Last November, Catherine H. van Zuylen, vice president of product marketing for Attensity Americas, a provider of business user applications from unstructured data, and Brian Koma, vice president of research strategies for Vovici, co-presented a webinar that focused on how marketers can use social media and online surveys to better target today's consumers. Here's a recap of that webinar.

Koma started by identifying today's new breed of customer, who uses the internet not just to learn what others think, but to share his own thoughts as well. These customers look to the web for solutions to their own problems with products and services, and may solve others’ problems as well by posting solutions that have worked for them. Today's consumer views the internet not just as resource to be read, but as a community in which to participate.

This has led to the rise of the social web — i.e., sites whose content is created by the users themselves. More than 100,000 suggestions, tips and tricks, for instance, are posted to expert forums by consumers every day. Looking at Twitter alone, more than 15,000 tweets are sent every minute.

Twenty percent of these tweets mention brand names, according to a study from Penn State University. Multiply this by the different blogging, microblogging and social network sites, and you have the potential for a tremendous amount of conversation taking place online about your brand.

That’s a lot of conversation taking place. So how do you make sense of it? Many people ask her if they really need text analysis, van Zuylen said, or if they can just get by doing it manually or using keyword counts.

The problem with manually reading all the comments in an online discussion thread is that while people start with good intentions, they usually don’t get through all the different discussions. Sometimes they just end up creating a word cloud or doing a keyword count. This may be fine for a one-time look at what the social web is discussing, but now think about trying to gauge the pulse of that every day, with new posts coming in hourly. That’s a lot of work. If you have a popular brand, you’d need to add staff to do this full time.

So marketers need some text analysis. Keyword counts are a simple analytical technique, but they ignore sentence structure. For instance, analyzing hotel reviews and tallying “room” and “smell” produce false positives like, “Although it was a smoking room, I couldn't smell anything. It seemed to be well-ventilated.” A keyword count would index that as a problem, when in fact it's a compliment. Text analysis, in contrast, parses sentences: ((The (room)) was clean) but (I smelled dog (in (the hallway.)))