For some time now, marketers have been wrestling with Big Data, the tens (and hun-dreds) of thou-sands of bits of customer intelligence gleaned from direct response campaigns, contact centers and other sources. This has become particularly important in building relationships with influencers. Social media has replaced the traditional influencer model with millions of mass influencers and opinionated customers. Billions of posts are shared every month on Facebook alone, and 25% of them mention brands. Marketers that tame the Big Data beast—that can aggregate, normalize, mine, analyze and act upon it—can achieve a huge return on investment from earned- and owned-media campaigns, and optimize their marketing mixes. In influencer marketing, there are at least two kinds of big Big Data to manage. One is aggregated influencer information. This comes from syndicated primary research consisting of profiles from public social media sites (Facebook, Flickr, Google+, LinkedIn, Twitter, YouTube and others). It also can come from raw customer data which, if you can unlock its stored potential, can identify attributes that can move markets. Being able to pivot on the data, to drill down—and sideways—to understand both context (relevant previous posts) and relationships (who is sharing content with whom) is crucial. The other kind of data to manage is real-time data, especially social media conversations among influencers and customers. For major brands, keyword searching can't keep up. One of our clients—an integrated marketer in the bedding industry—said: “My audience is anyone who wants to get a good night's sleep. [But] type "sleep,' "bedding' and "mattress' into a typical free-search tool and you get millions of results.” Technologists are trying to help marketers manage, if not master, the Big Data problem. Here are some of the ways they're attacking on multiple fronts:

Identifying key influencers in their markets. The most comprehensive databases—those used to contain the top outlets, journalists, analysts and bloggers—must now scale to accommodate millions more global digital influencers. And those profiles must be managed, refreshed or deleted to ensure their ongoing relevance.

Accurately measuring and rating true influence. This can be accomplished by using Web metrics, multifaceted criteria and primary research to determine who is exerting influence in specific markets. Reliable data let marketers segment their influencer bases (and build their A lists) and adjust priorities. But influence ranking is a dynamic metric; it changes continually and must account for such variables as actual subject area expertise or geographic location.

Aggregating content from hundreds of millions of sources. These include news sites, blogs, social media sites and traditional media.

Analyzing the content. Use resources such as sentiment analysis tools to professional consultants and data scientists.

Acting upon both the profiles and real-time data. Build relationships and engage with digital influencers, online communities and individual voices.

Two enabling technologies are key. Cloud computing offers the scalability, data storage capacity and significant cost reduction essential to Big Data mining and management. The other, natural language processing, will revolutionize search by replacing traditional, Boolean queries. By accurately recognizing relationships between words and sentences, natural language processing will return much more relevant search results. Peter Granat is president-COO of media intelligence and information services company Cision North America (http://us.cision.com). He can be reached at peter.granat@ cision.com.