The SMAC Stack™ (social, mobile, analytics and cloud) is what we call the "fifth wave" of IT—after mainframes, minicomputers, distributed PCs and the Internet. Its emergence is causing significant disruption in the business world, forcing organizations to update their IT architectures and data management systems to remain competitive.
They are impacting data management thusly:
Social networking: According to a recent Pew Research survey, 72% of online adults in the U.S. use social networking sites.1 All of this data — employment history and connections, relationship status, personal and professional
interests — can be mined for insights across three dimensions: customer behavior, networking potential and brand sentiment.
Mobile: Customers today expect to continuously engage with companies across multiple channels (mobile, social, Web, phone, in-store) and have a consistent experience. Businesses need to analyze real-time information on consumer location, behavior and needs to enable delivery of the right product or service, across all channels.
Analytics: Enabling decision-making through advanced analytics is critical to unearthing insights into customers, products and locations, using data from multiple sources, including unstructured data (both internal and external), structured data and third-party information.
Cloud: The vast volumes of data, varied content types and complex integration scenarios generated by social and mobile platforms require a real-time and integrated approach, utilizing a cloud infrastructure.
3-D Data Management
Traditional precepts of data management — such as approaching data quality as a linear function or consolidating data in a static data store — are inadequate for optimally utilizing big data and meeting demands for localized and personalized transactions across channels. We propose a new approach for organizing; transporting, analyzing and managing data assets, one that operates across three dimensions (see Figure 1).
Dimension 1: Integration
In order to deliver data expediently, economically and in a timely fashion throughout the business, we propose a five-layer data integration architecture we call the GRiD — short for Get Right Data (see Figure 2).
Understanding the GRiD
Dimension 2: Data Fidelity
Focus on quality, or fidelity, of data and the quality of the insights that can be generated must include governance (providing the right framework for the data), data volatility (dealing with the dynamism of data over time) and data integrity (ensuring accuracy, consistency, completeness and veracity).
We propose pursuing data quality efforts within larger operational processes rather than as a foundational element. This can be accomplished by establishing "process hubs," which include four components:
Define the end-to-end business process and sub-processes.
Identify the linkage between the process and the data (e.g., "create, read, update and delete," or CRUD), as well as the process and data pain areas.
Document process-related performance metrics and use these to establish data metrics.
Establish data governance by using the business rules surrounding the process.
By using process hubs in place of global hubs, business can realize several benefits: such as easier establishment of qualitative success metrics; fewer variances in data definitions and better boundaries around the scope of data management.
Dimension 3: Data Analytics
To control costs, obtain actionable insights and perform predictive analytics, we propose an approach already being taken by many forward-thinking companies: crowdsourcing, in which analytics models are built by talented individuals in the public domain.
Companies set up these projects as coding challenges or competitions and issue a monetary reward for the best submission. Industry leaders such as Netflix and Allstate are prime examples of how crowdsourcing can benefit business objectives at a fraction of traditional IT costs.2
In the future, additional dimensions may be added to the 3-D data management approach. We can foresee the value, for instance, in "social data management," in which organizations are able to identify influential individuals on social media and interact with these individuals to encourage specific patterns of behavior.
For real-life success stories of 3-D data management, please read our whitepaper 'Don't Let Your Data Get SMACked: Introducing 3-D Data Management' [PDF]. Please visit Cognizant's Enterprise Data Management Practice for more information.
2 "Allstate Announces Crowdsourcing Effort to Test Usage-Based Product," Allstate Web site, July 25, 2012, http://www.allstatenewsroom.com/releases/allstateannounces-crowdsourcing-effort-to-test-usage-based-insurance-product and "Learning from the Crowdsourcing Efforts at Netflix," CrowdSource, Sept. 11, 2012, http://www.crowdsource.com/blog/2012/09/learning-from-the-crowdsourcingefforts-at-netflix/.