By Dick Martin
Everyone’s talking about Big Data, even those of us with liberal-arts degrees. It’s suddenly dawned on us that the metaphorical “cloud” into which our digital devices exhaust is an actual orchard of data about our purchases, conversations, and movements, ready for round-the-clock harvesting.
Amazon alone catalogued every detail of the 426 sales per second it made at the height of the last Christmas shopping season. Meanwhile, the who, what, when, and where of our lives—from the measure of our creditworthiness to the age of our automobiles—accumulates on thousands of other corporate servers.
IBM calculates that 90 percent of all the data that ever existed was created in just the last two years. And it’s growing by 2.5 quintillion bytes every day. That’s roughly the storage capacity of 3.3 billion CDs. And it will increase exponentially over the coming years. Cisco forecasts that 50 billion devices and objects will be connected to the Internet by 2020. Plug something into an electrical outlet and it will report its usage to corporate headquarters when it’s not gossiping with all your other appliances.
For decades, that miasma of data was a cloudy soup of 1s and 0s—vast, unstructured, and lying just below the surface of usefulness. But the people who mine data have figured out how to extract meaning from them. In one famous case, Target knew a teenage girl was pregnant before her parents did, based solely on her shopping habits. Mom and Dad weren’t tipped off until discount coupons for maternity clothes started showing up in the mailbox addressed to her.
The promise of striking gigabyte gold prompted Harvard Business Review to declare “data scientist” the sexiest job of the twenty-first century. It’s more than a geeky coding job. Finding actionable insights in data requires a combination of right- and left-brain skills—computation and logic infused with curiosity, intuition, and imagination. PR people like to think they have a corner on the latter and can rent the former whenever they need it.
When I ran PR for AT&T about a decade ago, we thought we were pretty sophisticated. We commissioned the usual tracking studies and paid an outside firm to monitor media coverage of the company and our major competitors. But all that information came in binders that were about thirty days out of date when they landed in our in-baskets. To identify influencers, we had to count how many times news stories about us mentioned them. We did our long-range planning on a whiteboard in a conference room that doubled as a “war room” for occasional crises.
Today, the PR News Operations Center at AT&T’s Dallas headquarters monitors 150 million online sources to measure the sentiment of what’s being said about the company and its competitors in real time. The information displayed on screens around the room and accessible to staff anywhere in the world tracks the company’s share of voice by topic and geography. It identifies major trends and pinpoints key influencers, tracking the path of a single tweet through cyberspace if necessary.
Brad Burns, AT&T’s VP of global media relations, says the center is a “big-time canary in the coal mine,” warning the company of developing issues before they become a problem. “But the real value,” he says, “is that we can make targeting and messaging decisions based on actual data—what people are saying, how they feel about issues—rather than on gut instinct.”
It’s also a competitive tool. Not only can the company gauge the reaction to its own news releases—it can track how customers are reacting to a competitor’s announcements, providing useful intelligence for sales and media teams.
The system was developed for less than $1 million in partnership with the company’s IT department and draws all its input from commercial data services. But Burns declines to call it Big Data.
“I’ve learned to think in terms of stages. At best, we’re working with Medium Data,” he says. “We’ve learned a lot, like the best time to issue a news release on certain subjects. But there’s still a lot we can do, like closing the gap between what’s trending online and what we’re displaying in our stores, and finding a way to use this data to predict behavior so we can get smarter about things like messaging and even staffing and store inventories.”
As a practical matter, such deep mining of Big Data is currently beyond the capabilities of most PR departments or agencies. But a few are nibbling at the edges. Not surprisingly, IBM PR has bitten off more than most.
In a series of what the company calls “social jams,” IBM held online brainstorming sessions with tens of thousands of employees, customers, and partners around the world. The sessions weren’t flawless: When too many people logged on simultaneously, the system crashed; few of the participants’ comments built on each other; and a few cynical postings in the early hours of the first jam prompted a senior executive to suggest pulling the plug on the whole thing. But each jam produced more than a gigabyte of text, expressing people’s thoughts and feelings.
Initially, it was more like a library of overturned bookshelves than a thoughtful essay. Then IBM applied its text-mining technologies to the postings. Text mining is a largely unsung subset of data mining. It consists of combing through a large volume of digitized text to find meaningful patterns. Instead of searching for a few predetermined key words, as the AT&T PR operations-center system does, it analyzes all words, categorizing them by meaning. Then it reports the most significant, based on frequency of mention, sentiment, relevancy, and other criteria.
Text-mining technology is just entering adolescence, all hormones and pimples. Even the IBM jams required a narrowly focused question, several iterations of input, and human intervention to arrive at usable results. But the technique has great potential to reveal what really matters to your customers, how they feel about your organization compared to others, and the secret to earning their loyalty and advocacy.
That’s why Big Data will increasingly shape the targeting, content, and performance of future public relations, just as it will transform every other corporate function. And that presents two interrelated challenges for PR people.
First, they need to work with senior management in thinking through the ethical implications of gathering and using all this data. I worked at AT&T back when the Internet was just leaving the lab. The economic value of the user data it would produce was already obvious. Less obvious were the boundaries and procedures for storing, accessing, and using it. That ignited a debate every company needs to have. And however that debate is resolved, companies must be transparent about the data they collect, and why they collect it. Tiny-type legal mumbo-jumbo that no one reads won’t cut it.
Secondly, within the context of those policies, PR people need to get as comfortable with data as they are with words. They should be plumbing whatever data that they have—Small, Medium, or Big—for actionable insights. And the most obvious place to start is in the world of social media. More than 15 million brands and businesses have pages on Facebook. The chatter on those pages can be either background noise or a flash of insight, depending on one’s analytical skills. Chevrolet’s PR agency, for example, analyzed the brand’s social media “likes,” comments, and shares to figure out what topics, timing, and posting lengths generate the greatest engagement.
Unfortunately, too much of public relations is based on gut feelings or past practice. The challenge is to use data—not on the back end of a program in counting clips or measuring outcomes but, rather, on the front end in the development of strategies. The only way to do that is to ignore the buzz and get busy.