Subscribe / Unsubscribe Enewsletters | Login | Register

Pencil Banner

Social data doesn't have to be big data to be useful

Thor Olavsrud | Sept. 26, 2012
Mention 'social data' or 'sentiment' these days and a conversation about big data is sure to follow, but you don't necessarily need a Hadoop cluster to leverage unstructured data--sometimes all you need is Twitter's native Search API.

He notes that this doesn't require Hadoop clusters or expensive data visualization solutions, just Twitter's Search API and a competent community manager.

"For our needs, we're able to use the Twitter interface directly because we have very specific searches that we have pregenerated," he says. "For the most part, what helps with the process at this level is to have a competent community manager that's constantly looking at the feeds and making a human decision to interact with somebody."

One example involves an athlete who was expressing concerns about a cough before participating in an Ironman competition. Using that information, Cold-EEZE sent a care package to help the athlete overcome the cough ahead of the race.

Genuine Interactions Are Essential

The key, Braiker says, is to create real, genuine interactions.

Whether you're looking at Twitter, Facebook or some other social network, he recommends identifying the people who will be most receptive to your message and then engage them with quality conversations. Remember details about them and what they've said in the past. Use names and talk about things of substance.

"Sometimes companies make really big mistakes because they just start promoting themselves nonstop," he says. "It's almost like being on a date and not going through the process of trying to know someone first. It really undermines how those tools are best used. You have to genuinely care about their interests and create a true connection with real conversation."



Previous Page  1  2  3 

Sign up for CIO Asia eNewsletters.