Social media has become the world’s biggest pool of market research, but how is that data being collected and analyzed, and how does it ultimately improve the consumer experience?

With advances in Big Data, social media data-mining has evolved a great deal in the past five years thanks to contributions from web search metadata and heat maps that signal end user activity. New social platforms like Pinterest have emerged to indicate consumer intent, layering even more data for the contextualized profiles used by brand marketers. But as things get easier for brands, do they also get better for consumers?

SiliconANGLE recently talked with Jeff Revoy, President and CEO of Viralheat (acquired today by Cision), about the potential uses of social media for predictive analytics, how to understand what content is relevant and how such data findings ultimately impact consumers.

Data runs the kingdom

Q: How is Viralheat different?

Revoy: Our founders didn’t come from marketing, but from Big Data and scalable networking. Because we’re data oriented, we have the ability to do everything end-to-end on the social side — publishing, analytics, enterprise integration. Content is king, but data runs the kingdom. It ties into your interest around predictive [capabilities] and what can be done with social data. We solve the problem of social data and come up with some nuggets of analysis.

 .

Q: What are the proactive and predictive aspects of leveraging social media data?

Revoy: Brands spend enormous amounts of money for offline data just to get perspective on customers, when the online social world is the greatest pool (review sites like Yelp and Glassdoor). Most companies are only harnessing a small amount of that data. The idea of using it for audience identification or predictive analytics is still very, very early in terms of people understanding how to take full advantage of it for brands and consumers.

One side looks at brand affinity, the other side looks at definitive live changes where a person moves from one bucket to another – getting married, having a baby, graduating school. These actions can be indicators to brands.

An example: someone gets engaged – now their online friends become potential for other services to target (think registrations, travel).

 .

Q: Any usefulness in data around people fantasizing about these new buckets before they reach that point?

Jeff Revoy of Viralheat

Revoy: That’s just about scale and machine learning over time, and identifying true human intent. We have a patent around intent and sentiment because these are tough problems to solve and we have opened up APIs to developers to improve this. From my years at Yahoo!, Inc. [I’ve learned] that’s how search has improved. It’s amazing how far search has come.

For some instances you need Pinterest specifically, but this data can be useful. One example is a negative sentiment about marriage, and you look more closely and discover it’s a tweet where someone’s making a joke about marriage. Maybe that’s data a divorce attorney could use.

The information is valuable to brands either way, and consumers as well. An example is how my wife used my computer to make a purchase, and for two weeks I was followed by this brand completely irrelevant to me. The more consumers can share, the better the ads.

 .

Q: What other use cases stand out?

Revoy: We launched our enterprise version just over a year ago. We’ve had a lot of success with case studies using Viralheat within the enterprise. For instance, we’ve done a lot of work in entertainment space. As an example, we’re able to cross all social networks and see who interacts with specific television shows. We can see what other brands [social media users] actively engage with. We can show how brands can utilize their marketing resources.

One thing we look at for the entertainment sector is a heat map as to when the best times are for interacting with consumers for TV shows. This data gives brands times to target to maintain relevance, especially for a place like Twitter where data becomes irrelevant almost immediately.

 .

Q: So are practitioners the real Big Data success stories?

Revoy: I think so. I come from a search background, and the first phase was very typical of the hype phase. Low-hanging fruit for how companies use search. Look at those same teams now. They’re very data-driven. I think social will go through the same changes. If you can gain insight, you can deliver a compelling experience.

Modern horoscopes and mainstream adoption

Q: Working with social data, have you come across any indicator that a consumer would respond to a daily horoscope service based on personal data?

Revoy: Seems there’s a novelty in horoscopes. I know that particular writers with a certain amount of whimsy tend to capture my attention, even if I don’t heed their advice.

 .

Q: Will data-driven recommendations ever become a separate service, or will it be built into the business like Amazon’s model?

Revoy: Amazon is a great example of how a company uses data based on your actions. There’s opportunity for a business to give data back to consumers in many forms — recommended items, email newsletters — not entirely different from offline companies today. If we can find that data on the web, we can be more proactive and give them ads that are more relevant.

 .

Q: Will consumers heed to modern horoscopes as a recommendation service?

Revoy: I think of a political question: Do you trust elected officials to offer security? In a tech-driven world, will we realize we don’t even have the right tools to manage? The signal vs. noise problem has always been there. Consumers don’t change their behavior so much as tech makes info more accessible. It’s easier to point and click and get a product delivered in two days, and that retailer’s hack may not have even affected my credit card.

photo credit: kevin dooley via photopin cc