Because digital services are the gateway to everything we do, nearly all our activities are quantified. How do we get individual data streams talking to each other?
Will Turnage, VP Technology & Invention, Technology
Big Data is a big tease. It teases us with stories of automated networks that extract insights from untold amounts of data in real time. It teases us with promises of personal optimization and real-time business decisions based on emerging trends and behaviors in the marketplace. And it teases us with the notion that it’s all done with the push of a button.But for anyone who’s actually tried to implement big data across multiple touchpoints, he or she knows it’s not always a simple process.
Managing Personal Correlations
On a daily basis, someone could use Beddit to analyze his sleep behavior, Mint.com to track his spending habits, Foursquare to document where she goes, and FuelBand to measure her activity. We can learn what’s keeping this person up at night, how he manages his money, what places she’s likely to visit, and roughly how healthy she is, but we have no idea how his sleeping habits affect his spending habits or how her activity levels affect the places she visits.
• What if, on days a person gets eight hours of sleep, he spent 50 percent less than on days when he got only get six hours? Could Mint.com improve financial suggestions by accessing a user’s Beddit account?
Of course, that’s the tricky thing about data. Cause-and-effect relationships are difficult to prove, especially given the complexity (and near impossibility) of proving an absolute relationship. What’s possible, however, is discovering correlations between seemingly unrelated datasets, in the hopes of gaining a more complete understanding of ourselves. But we are not islands. There are a variety of factors that influence our health and behavior, such as weather, the stock market, or traffic congestion. This also applies to large organizations and the insights to be gained by analyzing proprietary data in conjunction with external data sources.
External Data Points and Unexpected ROI
The first step in unlocking data in seemingly unrelated datasets is to think like an investigative journalist: look for data in unexpected and new places for the greatest discoveries.
The largest potential return on investment (ROI) won’t come from analyzing existing company data on its own but complementing it with public data sources. In the past four years, we have seen more and more public data made available by local, state, and federal governments as well as numerous government agencies. This information used to be available only in print, but now that it’s available digitally, it provides private companies with an easy way to incorporate it into their own data and discover strategic insights about their business.
For example, the National Weather Service’s website provides access to historical weather data. So what, right? Wrong. Say, for example, a brick and mortar retailer used this seemingly inconsequential and unrelated data to compare daily sales in each of its stores with the actual weather for that day. The retailer could then answer questions like: On rainy days, are sales higher or lower than on nonrainy days? The retailer could see that sales always spike during the first cold snap of each fall season. Equipped with these new insights, the retailer could optimize store offerings based on weather patterns, which would impact future organizational strategy.
Of course, companies need to be able to find public data in order to use it. New tech companies like Enigma.io are building business models around helping consumers access this publicly available data. Enigma.io is a NYC-based company that connects thousands of disparate government datasets, exposing in the process information that users might not know even existed. For instance, searching the Enigma.io portal for the word McDonald’s reveals that each McDonald’s location has an FCC license to operate its drive-thru window. So to track the growth of McDonald’s, anyone can just look at open public government data.
Still not convinced of the ROI from analyzing external data streams? McKinsey estimates that open data can help unlock $3 to $5 trillion in economic value annually. Companies who are willing to look for data in unexpected places will reap the most benefit—for both their bottom line and their customers.
On the hunt for unexpected data? Check out Jess Greenwood’s piece on mindful data, which explains why bigger isn’t always better.