If the super-charged computers that Boulder-based company Umbria uses to break down the blogs, message boards and online social networks into lucrative market research had feelings, they’d be blushing right about now. Over the past few days the blogosphere – which the company number crunches into a massive, constantly-updated focus group for the benefit of its well-paying corporate clients – has been buzzing about Umbria.
If you like this story, consider signing up for our email newsletters.
SHOW ME HOW
You have successfully signed up for your selected newsletter(s) - please keep an eye on your mailbox, we're movin' in!
On April 2, J.D. Power and Associates, a massive global marketing information company, acquired the plucky little blog-searching company. The development caused Umbria’s chief technology officer, Ted Kremer, to be even more vivacious than usual. “The opportunities gained from joining with a company that has 40 years of marketing research experience and understands market research research better than anybody are tremendous,” he says. “It's exciting in that Umbria will now have the brand power of JD Power to go with the technological power we have developed over the past four years. Their understanding of how customers impact product design and product rollout, combined with our ability to passively listen to what people think, is a really powerful combination.”
In other news, says Kremer, cosí, Umbria’s search and discovery tool, is “doing phenomenal.” Using complex data-mining and natural language processing techniques, cosí has long been able to “read” the blogosphere (or the closest thing a computer can get to uniquely human process of reading). Unlike search engines, which must identify the most relevant mentions of a particular topic, cosí must find every mention of that topic and categorize each one – is the mention positive, negative or ambivilant? – and provide insight into its author, such as their gender and approximate age. Lately, however, it’s learned new tricks, like searching for particular types of authors. “Before we always started with a topic,” says Kremer. “Now we can search for a set of characteristics of the authors themselves.”
For example, Kremer says he can instruct cosí to “build a tribe of Hillary Clinton supporters and John McCain supporters” by seeking out blog entries that the computers determine to be penned by these two factions. “Now we can compare and contrast these two groups of people. What do they have in common?”
Of course, we don’t need super computers to tell us the answer. They’re both scared of Obama. – Joel Warner