FORTUNE — A spate of acquisitions of big data startups has drawn attention to an already hot category. In December, Apple (AAPL) bought Topsy, a social data analytics company, for $200 million. Last week Google (GOOG) outbid Facebook (FB) for DeepMind Technologies, an artificial intelligence and big data company, paying $400 million, and last year, Google paid just under $1 billion for Waze, the crowdsourced traffic data app. Adobe, WPP, and others are reportedly bidding for an Indian social analytics startup called Simplify360. This week Twitter (TWTR) revealed it made $23 million last quarter on data licensing; earlier this week Foursquare struck a deal with Microsoft (MSFT) to allow the conglomerate to license its location data.
Big data is a big business, even if everyone working in the field actually hates the term. Some argue it’s too broad, others say it’s purposely all-encompassing. Chris Moody, CEO of Boulder, Colo.-based Gnip, says the term big data incorrectly focuses on the size of the data set, rather than what’s most important: the quality.
He should know: During the Super Bowl, Gnip processed more than 90 billion “evaluations” per second on pieces of data. Gnip powers the social media data sets for hundreds of companies, from Adobe (ADBE) and Salesforce (CRM) to Klout and Union Metrics. It’s one of three companies in the world to have access to the full Twitter firehose. (Without a partnership, Twitter only makes a portion of its social data available.)
The “big” part is impressive in terms of sheer volume — Gnip pulls data from Twitter, Tumblr, WordPress, YouTube, Facebook, Instagram, Foursquare, and others — but all those terabytes of data mean nothing without a filter to give them some context. Gnip’s secret sauce is the way it filters its social data into something meaningful to its clients. The company has been able to detect where earthquakes have occurred faster than seismic radars, glean insights on how cholera has spread through Haiti, track interesting patterns in how Iranian politicians communicate, and predict stock movements for hedge funds. Earlier this week, Gnip announced that, alongside Twitter, the company would provide “data grants” to non-commercial academic researchers, who will be allowed access to their data for free for their projects.
But most important for Gnip’s business clients is its ability to unearth a “million-dollar Tweet” with its filters. Examples of such a valuable find include a Tweet that reveals the location of a new Wal-Mart before any public permits are filed, or a Tweet that leads to a stockbroker trading a stock. During the Super Bowl, for example, Gnip was looking for “90 billion needles in a haystack on behalf of our customers,” Moody says.
The best example Moody gives is that of customer engagement, which is the most common use of social analytics and data. Over the Super Bowl, Adobe blended data from Gnip with its own data to help brands understand how much of a lift their ads achieved. (RadioShack won, with a 22x increase in brand mentions around its ad.) The company did the same for predicting box office sales for movies. Before the rise of social data analysis, brands would get feedback on their campaigns and products through surveys, focus groups, and market research, which takes months. Big data makes it immediate. “It’s very easy for us to plug into their real-time infrastructure and combine it with our data infrastructure for behavioral analytics and deliver it to brands that’s consistent with that they want,” says Bill Ingram, VP of Analytics and Social Solutions for Adobe. “It’s a very elegant real-time solution.”
Launched in 2008, Gnip has only raised $6.64 million in venture capital, most recently in 2010 from Foundry Group and First Round Capital. With 85 employees, it’s safe to say the company is now financing its own growth. Such a small raise makes Gnip an anomaly in the enterprise tech world, where companies must often raise hundreds of millions of dollars across many rounds of funding.
Others operating in the social media data category have raised large rounds of funding: DataSift has $71.7 million in venture backing. Dataminr, which recently partnered with Twitter and CNN to create a tool for journalists, has raised $47.6 million. Link-tracking site Bitly has raised $28.5 million. Social data site Gigya has raised $69.8 million in venture funding.
Gnip’s small raise is not for lack of interest: Gnip is often approached by venture capitalists once they learn how many startups are powered by Gnip’s data, Moody says. But the company has no plans to raise additional capital at this time, he notes.
The initial value of Gnip, whose name is the word “ping” spelled backward, was that it cut down on the number of data requests to Twitter from third parties, which was a big problem back in the “Fail Whale” days when Twitter outages happened frequently. Now that Twitter is stabilized and Gnip offers social data from many sources, the company finds itself playing an integral role as a broker and synthesizer of social media signals. Regardless of whether size really matters, as a business, big data is going to be huge.