BrandAura is a just-launched social media analysis tool package that’s planning to be an affordable choice for marketers and branding analysts to provide clients with insight into what’s being said across the web about their clients’ products or services or trends around them. It’s the brainchild of Dr. Andrew Starkey of the University of Aberdeen and its spin-out company, BlueFlow Ltd. http://www.brand-aura.com/.
So, how does this fit into the world of sentiment analysis solutions? “Sentiment analytics is essentially a list of words and you compare the scores of how these lists of words do. That’s the antithesis of us,” Starkey contends. “We work out the context first, what is said about particular topics by a community. You solve that first and apply a sentiment analytics layer after and then you get a better picture of what is being said and a more accurate picture.” The core, he says, is about language independence and about what people are saying in context.
The engine that drives the contextualization analysis is not using AI or NLP, which Starkey refers to as “black boxes.” Instead, “it is a heuristic which tries to create a tree of information about the context of the words, in some respects similar to regression trees. The advantage is you can put AI on top or sentiment analysis on top so you get more layers through analysis and insight.” Meantime, he says, speedier outcomes are the result of not relying on technologies such as NLP, which requires breaking a sentence into its component parts and tagging the pieces. “We don’t need to do that because of the way we try to hone in on the structure of data within the whole data set — that is all comments made by everyone.”
The context and relevance of conversations can be visualized in a context cloud with BrandAura, where every word can be analysed within its own context. Clicking on a word puts it at the center of the cloud and you can continue drilling individually into each word. He notes as an example some analysis BlueFlow did for a whiskey brand in Scotland, the birthplace of the liquor that now is seeing more competition from companies in places like India. Scouring blogs from America, the U.K. and the Far East, it created a cloud with the Glasgow producer’s brand in the center, and closest around it were words like “gun” and “barrel,” which, without the contextual relevance its solution enables, could have been misconstrued if not missed altogether, as those words’ associations with the whiskey industry wouldn’t necessarily be obvious in common usage. As it turns out, he says, the company got excited because the tool recognized those words in context for its industryâ€”that some of their whiskeys are produced in gun barrels, and that those words’ close association with their brand in the word cloud shows that that has gotten the attention of the audience they’re catering to. “I think that’s essentially the difference between perhaps what we can do in having the context come through,” he says.
Self-Service Portal Available
U.K. based Emiore, a small digital data analyst shop run by Jayne Coulthard, partnered up with Starkey early on in working on and testing the technology, and applying it her showcase site of analysis as it relates to U.K. politics at Election Trends and her e-consultancy business. She’s also the digital marketing director for BlueFlow. “A lot of companies offer sentiment or analytics programs, but we offer something more sophisticated in terms of the depth of analysis. We’re bringing the context side of things into the picture â€“the context of each different word that comes out in relation to each otherâ€”and showing trends over time with context, so that’s appealing to those who want to look at business development, where the opportunities are in the market because we can dig deeper to what’s being said,” she says. “A lot of other tools are language processing and NLP, but NLP isn’t that great at finding things in context. That’s our advantage in digging through the data.”
Emiore â€“ and other such consultancies â€“ can use the BrandAura technology as the underpinning of their more comprehensive services, but those that just want a quick take at analyzing their data themselves can visit analysemydata. It takes advantage of BrandAura’s full automation capabilities for insight, sans the “bangs and whistles,” as Starkey refers to them. “For smaller companies that don’t have high-end customers but need a little analysis, the do-it-yourself module’s competitive advantage is that it’s 100 percent automatic,” Coulthard notes. “That opens up the whole contextual and data analysis option to a wider market, to people who haven’t been able to afford it before or who haven’t had clients with large budgets.”
And the speed at which it can process data and find connections in the word cloud lets her focus on bringing value-added services to bigger outfits with more substantial budgets, she says. “We publish the tools so people can automate things,” says Starkey. But if you want to follow the findings up in more detail and examine the themes that come through that, that is where the added value [of services like Emiore] comes in.”
• Don’t forget to propose your startup for our Semantic Web Impact Awards. The deadline is Sept. 15.