Steps to Big Data Success

By Sudhi Sinha, Vice President of Product Development, Building Technology & Services, Johnson Controls

A great convergence occurred several years ago in information technology, with the cost and capacity of computer storage, processing, and networking all improving at the same time. This trifecta brought us to an inflection point where new technologies such as Hadoop could unlock unprecedented volumes of data and deliver new business value. That is the essence of the much-talked-about phenomenon known as “big data.”

In practical terms, big data means it is now feasible to collect, store, and use all available data, leading to new insights derived from ever-more sophisticated analytics. This presents a great opportunity to increase the value of existing data while also preparing the ground for the deluge of additional data expected to be unleashed by the Internet of Things, or IoT— the network of information generated by people, objects, and the environment.

How did we get to this place? Historically, large data sets needed to be structured and stored in a very specific way in order to then apply queries and analytics. From the columns and rows of what conceptually was much like a vast spreadsheet, deriving insights was an inflexible process where the very structure of the data was predefined and tended to make the range of conclusions all but inevitable.  In that environment, out-of-the-box thinking was almost impossible.

Today, big data technologies have overcome the limitations of past practices by breaking data sets into multiple components, distributing and analyzing those components on different processors, and then aggregating the results to deliver a meaningful picture of the whole that is not constrained by a specific, preordained order or structure.

As the ability to use all that data and run analytics on it has increased and improved, businesses are suddenly able to access new types of insights. It is the “art of the possible” for data.  Furthermore, now that the analysis encompasses all the data rather than just samples, the results are more accurate and   you can look at data in connection with more scenarios, sources, geographies, and applications. Your view of what is possible starts to broaden.

Combining Big Data and the IoT

Big data technologies are maturing, and they are widely applied across many industries. In automotives, for example, a manufacturer can incorporate data from warranty and maintenance activities or even a customer’s visit to the auto body shop to better understand patterns of failure.

Collecting data from dealers in different states is not new. But big data takes this activity to the next level. For example, marrying the actual weather data in a region to vehicle failure information allows manufacturers to begin developing predictive analytics. Somewhat simplistically, cars in Massachusetts may be found to suffer more cold-induced battery failures while vehicles in Texas may see more problems with air-conditioners.  But folding in other data streams flowing from the Internet of Things enables more nuanced predictions. Think how the somewhat obvious prediction that cars in cold climates suffer from battery failures and in hot climates experience air-conditioner problems can be enhanced by factoring in local traffic information, say, or sensor data from the car itself.

Consumer data is central to retail operations and to the airline, ticket, and travel industries. In these instances, you are trying to understand demographics and individual consumers in the context of location, season, trends, and so on. Big data is a powerful tool for bringing together what you know about those factors and integrating new sources of information – for example, sensors in a retail facility or weather for a travel site – to produce deeper insights.

Putting Big Data to Work in Your Business

To realize the value of big data and capitalize on the emerging  Internet of Things, follow this step-by-step approach:

1. Build a Strategy Framework.
  It is vital to develop a clear conceptual understanding of big data and how its potential might map to your own organization’s needs.  This should probably begin with a review of existing business intelligence efforts and data sources. But don’t stop there; look further and see whether there might be other data sources that aren’t being utilized or processes that could be better instrumented to provide valuable data. Remember, big data means BIG. More is generally better.  Think about the kind of insights big data might be able to provide, and then get ready for the next steps, when you will move ahead and engage the organization.

2. Create an “Opportunity Landscape.” 
 If the full potential of big data is a gold mine, the tactical projects that can get you started are gold coins.  The point is to focus initially on a small number of projects or initiatives with the potential for significant payback.  These could be aspects of your business that have not performed as expected and where data is available to potentially change the game.  When you’ve found the gold in those projects, you will be better able to gather resources to finance the gold mine and begin to extract widespread rewards for the organization.

3. Effectively Manage Big Data Projects. 
This means having a good grasp on the learning that may be transferable from other transformation and IT projects, as well as focusing on what is really unique.  Depending on the existing capabilities of your organization, big data may require investments in human capital—the IT experts and data scientists who can help ensure that you get the desired results. Because big data is by definition an evolving activity, having a robust project-management framework can help you keep things on track and moving in the right direction. That approach can also help you steer the initiative toward new and emerging opportunities.

4. Build the Right Technology Landscape. Even if you are an IT professional, the nuances associated with big data may be eye opening.  A big data initiative does not have to “break the bank” but it will likely require some specific investments.  The good news is that big data initiatives are generally less expensive and less complex than the massive data-warehouse projects that some organizations have built using traditional IT tools—a sort of “brute force” attempt to garner more value from corporate data. The best news is that starting small means your initial expenses can be quite low.

5. Build a Winning Team.
 Even in an era where technology is so critical, the human factor can make the difference between success and failure. Big data projects call upon a range of skills. Obviously, some of these are pure IT skills; others have to do with mastering the data science side of big data. But deep business knowledge is also vital – having the ability to dig into operational realities and discern issues that can be attacked with big data. Recruiting the people you will need for your big data projects, and then organizing, managing, and motivating them, is a case of making investments up front that will pay over the long term and yield future success.

6. Manage Your Investments and the Monetization of Your Data.  
The valuation and monetization of data is almost a “secret sauce” aspect of big data.  Having an approach to help you weigh the costs of big data with its benefits can help you make decisions that are most likely to be profitable and meaningful. This inquiry starts at a granular level and helps you develop an appreciation of the true value of data and its costs.  If you take this path you’ll never look at gigabytes the same way again, and you’ll begin to develop a practitioner’s instincts for harnessing information cost-effectively.

7. Effectively Drive Change.
 Your initial big data steps may not rock the boat too much, but over time, big data can upend assumptions, sometimes even assumptions that have been driving the entire business.  With big data, you have the potential to cut through the fog and know the things that have always seemed unknowable.  If that sounds hard to believe, think about how much data is not used effectively now. Add to that the new data that may become available in the near future, through sensors, better use of existing systems, and new external sources. To ensure that all of this data helps the organization, plan to implement change-management practices and start teaching your organization how to be more agile, adapting quickly to new assumptions driven by fresh insights.

8. Communicate Effectively.  
Change can only be mastered with effective communication. This isn’t simply a one-way street.  Listening is vital, too.  For an organization to ride the big data wave successfully, everyone should be “on board.” Only a small percentage of the organization may need new skills, but because big data can alter how business is conducted, everyone needs to be aware that change is afoot and to understand their stake in the future.

***

In the final analysis, big data is about opportunity. It is the embodiment of the old adage, “knowledge is power.” By rebuilding your business upon the collection and analysis of big data – data that will expand exponentially as the Internet of Things emerges and matures – you can create a better future.