Data: The New Oil? [Takeaways from TiECON East]

On May 30, I attended the annual TiECON East (#TiECONEast) conference in Cambridge, MA. The conference was organized by “TiE” (The Indus Entrepreneurs), which describes itself as the largest, global, not-for-profit organization promoting entrepreneurship. This year’s conference, themed “Breaking Boundaries”, featured technical panel discussions on timely topics ranging from the Internet of Things to robotics, Cloud security, and connected healthcare. One panel, entitled, “Data is the New Oil”, dealt with the question of how to “refine” and monetize Big Data. Like oil, data is a hugely valuable resource. But unlike oil, the ability to extract value from data is not very well understood, and a vast amount of data remains “untapped” (or, to put the latter point another way, at the risk of overextending the analogy – we will never reach “peak data”). Questions were posed to five panelists – including a panelist from EMC Big Data spinout, Pivotal. Here are the key points I came away with: 1) How can Big Data be sold to business unit executives?

• The “prep for the future” argument doesn’t work: Business executives have many priorities competing for budget and attention, and most of those priorities are near-term. Even though they may “get” the message that their competitors might gain an advantage using Big Data, preparing for the future is always something that can (and usually does) wait for tomorrow.
• Listen for the pain points: Many execs will tell you their Big Data challenges without necessarily realizing that they are. A mobile executive talking about subscriber churn; A utility exec talking about electricity theft; A logistics exec talking about fleet fuel costs — these execs might not view these as problems that Big Data can solve, but as it turns out, analytics is being used today to address all of these issues. It’s our job to listen and to think about how Big Data can be applied.
• In some industries regulatory compliance can be a key driver: Understand how analytics can be used to help CFOs meet the regulatory requirements in their industries (e.g., HIPPA, Sarbanes-Oxley, Basel, etc.).
• In most industries, incremental growth is the main driver: Whether it’s saving costs through efficiency or penetrating untapped markets with new products and services, the most exciting opportunities are the ones that promise to introduce whole new sources of top- and bottom-line growth.

2) Do companies have Big Data, or just big problems with the data they have?

• Many companies are grappling with “analysis paralysis”. A familiar example is a utility company trying to figure out what to do with the avalanche of new data being generated by newly deployed smart meters. Another example, cited in the panel discussion, is a mobile operator trying to identify problem apps on mobile devices in order to decrease support center calls.
• It is our job to help these companies discover the use cases. Look at the short-, medium-, and long-term value that a company can extract from its data. To use the utility example, a near-term use case might be to simply understand daily and seasonal electricity usage patterns among its subscriber base. A medium-term use case might be to use that new understanding to create energy efficiency incentive programs for subscribers. And a long-term use case might be to work with city planners to design automation into buildings, enabling them to adapt usage behavior to minimize the CO2 output of a regional grid.

3) What are some Big Data do’s and don’ts for IT?

• Don’t: Try to be a futurist. Do: Focus on agility:
• Trying to read the minds of business stakeholders isn’t going to deliver the Big Data solutions they need. Instead, use the tools available to “fail fast” and iterate. Faced with such a huge volume, variety and velocity of data, trying out many combinations quickly to see what works is the only way to deliver on the fourth v of Big Data: value.

• Don’t: Be a “middleman” between your users and their data. Do: Maximize speed:
• A common time-consuming task in traditional BI environments is the movement of data from one data repository (usually a production repository) to another so that it can be analyzed. Tools and platforms now exist to obviate this need. With in-memory data fabrics that enable real-time data analysis, and storage platforms that allow a single “data lake” to be accessed using different protocols, including HDFS, it is increasingly possible to analyze data in place and return results much more quickly.

• Don’t: Think all data is equal or that security is all about authentication and encryption. Do: Be aware of the unique traits of your data, and your users.
• A good example here is geofencing of data. For a variety of reasons, ranging from privacy to regulatory to performance, you might wish to store data in a particular region, country or datacenter. Another example is understanding the access patterns of your users. An increasing number of security breaches today are “inside jobs” (think of Edward Snowden). Security is no longer just about ensuring that untrusted people cannot access your systems. Security is also about ensuring that trusted people don’t access data they shouldn’t. Can your systems spot anomalies such as a user accessing a database multiple times in one day that she or he has never accessed before? If not, your data is at risk.

4) If Big Data is the “new oil”, then automation is the “new plastic”.

• Automation is where the true power of analytics, particularly real-time analytics, comes into play. It is commonly agreed that “data scientist” is one of the hottest new fields. The belief is that someone with a unique set of skills, combining statistics with depth of expertise in a particular business domain, is needed to separate the wheat from the chaff in business data and generate real insight. While this is true, it is also true in many domains that nobody has the time to do analytics on all the data coming in. There’s just too much of it. And furthermore, some decisions need to be made instantaneously (you wouldn’t, for instance, want a data scientist making braking decisions for your self-driving car!).
• Big Data problems are becoming “Fast Data” problems. And these problems are being addressed by technologies such as Gemfire from Pivotal, which provide the tools necessary to build decision models into memory so that actions can be taken instantaneously. But it is very early days for this kind of technology, and there was agreement on the panel that automated decision making represents the next wave of Big Data innovation.

A lot of good stuff here to think about for entrepreneurs and big tech companies alike. And judging by the energy in the standing-room-only room at this session, I’d predict that we’re going to be seeing an enormous amount of innovation in this space! (P.S. My apologies for the “new plastic” analogy. That was mine…)