Strive to be Data-Driven Rather Than ‘Numbers-Obsessed’

data-driven
Strive to be Data-Driven

One of the key assumptions about gathering big data is “the more data we have, the higher the likelihood will be well find intelligent patterns.” But this is all for naught “if we have no idea of what to do with it.”

That’s the view of Lutz Finger, director of data science and data engineering for LinkedIn and author of Ask, Measure, Learn: Using Social Media Analytics to Understand and Influence Customer Behavior.  Finger states that many big data problems can be toned down to actionable and digestible insights with a deceptively simple approach: “ask, measure and learn.” This formula – which many organizations overlook in their rush to become data-driven – helps realize value from even the most complicated big data analytics problems. The problem is, not enough attention is paid to the “ask” part of the equation, he said in a recent BrightTALK webcast.

http://www.amazon.com/Ask-Measure-Learn-Analytics-Understand/dp/1449336752

The importance and success of big data analytics all comes down to pattern recognition, Finger explains. “What’s a pattern? If you see a sandwich with a piece of chicken on it, and that chicken looks green, well you have a pattern recognition,” he quips. Figuring out where traffic bottlenecks occur every morning is another form of simple pattern recognition, he adds. However, for complex systems, such as customer buying patterns, the task of pattern recognition becomes more difficult, and more dependent on big data. Organizations recognize this, and they are identifying and assembling vast stores of big data.

The problem is that big data is essentially useless in most situations. “If have this problem, let’s throw all kinds of data on it, in the hope that this data is creating some kind of metrics, some kind of pattern, which I can use,” Finger says. “Not only did I not think about the question. But I didn’t think about the data either. You have to be specific about what you want to do.”

As a result, he continues, “suddenly I’m sitting on a bunch of data… the question always come down to what are you doing with the information?” he asks. “Big data is not always useful – just having it is not enough. What you want is actionable insights.”

The risk is “if you have data and really don’t know what to do with it you become numbers-obsessed,” Finger says, adding that “people use data to trade beautiful dashboards. But beware, there is a difference between data-driven and numbers-obsessed.” Many data scientists and data analysts may find big data problems to be rewarding and challenging. However, people on the business side don’t necessarily share this view, viewing big data as “big pain.”

Even “personal big data” can be challenging, he says. “You have made a lot of photos with your mobile, webcam, with your camera, and so on and so forth. If you look back one year, and you want to figure out where the photo of your daughter is riding her bike. ‘Ooh, I don’t know, somewhere on my multiple hard drives, or maybe in Google Drive or maybe in Evernote…. I don’t know.’”

That’s the gist of the personal big data problem, Finger continues. “You don’t know how to get to your data, and you don’t have a semi structured layer on top of the data. This also applies to business cards and all the other digital information you have on people. The challenge is converging them into something that is actionable, and becomes easy to use if you want to get a certain photo, or if you want personal information to contact. You need a way to get there.”

The bottom line – for personal big data as well as corporate big data, is “we don’t want big data, what we want is insights,” says Finger. “We want actionable insights, we want small information. The whole point of data… is to figure out how to build big data into small data.”

That’s why Finger helped develop a framework comprised of three parts: ask, measure and learn, which can be applied to data handling at all levels of organizations. “The model is quite simple,” he says. “Ask the right questions, then measure the right things, then learn from it.”

The hardest part of this process is the ‘ask’ part, Finger points out. “We’re so busy with technology, and so busy acquiring data, we never ask, and thus annoyed we never learn.” It’s asking the right questions up front that ultimately delivers actionable information, he says. This is a priority at LinkedIn, where the priority is to connect members with live contacts and potential job or consulting opportunities. “Data by itself is useless. Information is useless. What only counts is action,” he says. “While we have heard all that before, it takes a lot of effort – the idea of ‘why are we doing this now’ gets lost in between.”

To illustrate, asking the right questions is what enabled Google to overtake the search engine leader AltaVista in the early 2000s. While AltaVista promoted the fact that it was providing access to huge volumes of information, Google zeroed in on addressing specific queries. “Google asked the right question,” Finger recounted. “People aren’t looking for the most data, they don’t look for ‘big data’ for themselves. They’re looking for a solution.”