INNOVATION
7:54 am
Tue December 18, 2012

How Retailers Are Mining Our Emotions

The Holiday Shopping Mood Meter, which tracks shopper sentiment at major stores.
Credit Netbase

How Retailers Are Mining Our Emotions

At the South Shore Plaza in Braintree, shoppers roam through the mall with shopping bags in one hand and a smartphone in the other. Welcome to holiday shopping 2012, when a good experience is shared online, and a bad one can go viral in a click of a tweet. But what you may not be thinking about as soon you hit “send,” is where those posts end up.

Natural language processing developer Netbase is mining them out in Silicon Valley and spending a lot of time trying to figure out what consumers are thinking, and applying those posts to their Holiday Shopping Mood Meter.

“We not just understand when someone is talking, we also understand what they’re saying, and we understand the sentiment,” said Lisa Rosner, Chief Marketing Officer of Netbase.

Mining the Internet for feelings

"What the mood meter does, it captures the conversation across the whole social web. This is not just Twitter. It’s Twitter and FB, YouTube and news comments and blog posts and any conversations that happen across the social web,” Rosner said.

Here’s how it works: when you tweet or blog about that “amazing” green sweater you found at Macy’s, that translates into a positive, and puts a spike in the mood meter.  On the flip side, if you absolutely hate a product or had a bad experience with a sale person, that store’s rating goes down.

"And then we filter a score, which goes from negative 100 to positive 100 and it really captures how people are feeling about the brand,” said Rosner.

“Feeling” is the magic word here. For years, the web has come up with ways to learn what we’re saying. Netbase takes it to a psychological level. With a staff of fourteen Ph.Ds in computational linguistics, they spend every day deciphering and encoding slang and abbreviations, like how the letters “gr” and number “8” equals “great”. Sounds simple enough for a computer to figure out, but not as easy when decoding the nuances of language.

"The colloquialisms are much easier than things like sarcasm, so we know with a high degree of accuracy when someone says that something is “sick”, they’re either talking about the flu or something being the coolest thing in the world,” Rosner said. “But where it gets complicated is when someone says, ‘I love root canals.’”

Computers Don't Understand Irony Because They Haven't Lived, Man...

Sarcasm and irony are tricky things for an algorithm to decipher, so the Holiday Shopping Mood Meter is about 80% accurate. The mood meter is tracking our consumer sentiment for ten retailers, including Amazon, Target and Home Depot, and it’s updated every hour. So what do retailers do with this information?

“The application of understanding this data can go from customer service to advertising creative to even things like where to place things on the shelf in the stores,” Rosner said. “We have customers who figured out to place the cheese next to the eggs instead of next to the meat because people talk more about cheese omelets than cheeseburgers.”

While the intent seems all and well, doesn’t it feel a little like Big Brother is trying to be our friend? Jay Baer, President of Convince and Convert, says yes, that is what’s going on here. Baer consults big brands on how to market themselves using social media.

“There is a strategic eavesdropping element to all of this,” said Baer. “And that is massively powerful for retailers. But it also requires consumers to understand that inherently you are what you publish and the Internet never forgets. For hundreds of years businesses have said wouldn’t it be amazing if we could eavesdrop on what customers are saying? Wouldn’t that be a powerful opportunity? And today that opportunity is a reality. It’s as if you had your own little patriot act. Or your own little ability to read everyone’s e-mail,” Baer said.

For All Your Complaints, Just Tweet

Baer says more companies are turning to the social web to find out what we think of them. But he says we, the consumers, are doing the same thing.

"What we find is that increasingly, consumers who complain about a brand in social media expect that brand to respond. Increasingly, consumers think of social media as the new telephone. And that is something that companies of all sizes need to understand. The Internet has made us all massively passive-aggressive, and so increasingly somebody just Tweets and says, ‘these guys are a bad company. Don’t support them,’” said Baer.

That is exactly what retailers don’t want this season. They want us all to love them, and whether or not you think it’s creepy or fair game, one thing is for certain: they are listening and watching, and taking note.