Big Data Analytics: Roadmap for an Enterprise

Authored by Gopal Appuswami, Service Line Leader for Payment Analytics, Opus Consulting

Big Data is transforming today’s analytics world across all industries. Big Data analytics is going to create and sustain competitive advantage for the companies of the future. It will change the way how businesses and IT leaders can develop and manage the related business processes and technologies. One of the crucial questions to answer in this relation is how does a company build an analytics roadmap using Big Data as the key theme. A roadmap is always a good asset to have before embarking on any journey through an unfamiliar terrain. So here are some recommendations on how companies can go about building this roadmap.
Avoid the Hazards Associated with the Hype
The first step is to learn to distinguish between the actual potential from the extravagant claims of what big data can do. Much of the hype rests on the following false promises:
1.    Big Data technology will identify business opportunities all by itself
2.    Harvesting more data will automatically generate more value
3.    “One size fits all” – the data in one use-case is relevant for all use cases
We therefore need to identify and avoid some of the “road hazards” that organizations often encounter in their eagerness to “harness Big Data for advanced analytics.”
Big Data programs can become really ‘heavy’ as a result of incorrect attempts to capture, store and analyze greater volumes, varieties and velocities of data because - by definition that is what Big Data is supposed to do. Organizations cannot appreciate precisely where the data insights are hiding. A common assumption is that an exhaustive analytic approach that processes all the data in the enterprise is the best way of unearthing value. This produces some unintended consequences, like:
•    Extensive efforts spent on moving data from one warehouse to another;
•    Delays in the discovery of the intelligence needed to take action;
•    Applying the same analytics to every business use case and problem;
•     And discontinuity between inputs and the desired business outcomes.
There’s no perfect Big Data Analytics Toolset
Another “hype hazard” to avoid is the unending search for the perfect “Big Data Analytics Toolset.” One Financial Services company and their IT service provider got carried away evaluating tools - any new tool that the technology team heard of was snapped up, and pilot projects were undertaken. The intentions were good: identify the toolset for what would be a production-grade, Big Data solution that could yield the necessary business benefits to ease their reporting needs, minimize the risk of conducting business, and/or identify new opportunities within existing businesses. After many months of this repetitive exercise, the company had so many pilots and so much data in silos that they had to scrap the entire initiative and engage another vendor to implement a more result-oriented “advanced analytics” capability.
Start Small to Manage Big Data
Companies that successfully harness the power of Big Data tend to start by applying advanced analytics to solve a small number of high-value business problems with existing in-house data and talent before investing in newer technology and talent. Hence, rather than collecting more data, and spending more money and time managing it, they use their existing enterprise data in a more intelligent way.
A gradual transition to what we call the SCALETM methodology (Smart, Clean, Accessible, Lean and Extensible) is an approach to managing big data in a small way. It is focused on sets of data that deliver specific business outcomes. By taking this approach, institutions can withstand the increasing internal and external pressures faced by their competition and look forward to a future in which they listen, learn and engage better with their customers.
What does all this mean? Here are some simple strategies to unearth the potential of Big Data using the SCALETM model.
•  mart: Smartly assess the business case, associated data, and define goals that are customer focused. Ensure data availability and feasibility to apply relevant analytics to your consumer base. Agree on the business objectives to be accomplished.
•  lean: Clean and identify relevant data assets that qualify for the business outcomes. Break down silos, challenge conventional wisdom, and consult multiple stake holders. Identify and assess the risk factors. Lay the foundation for business analytics using big data.
•    ccessible: Configure the processing steps needed to collect, store, analyze and visualize the data. Translate hypothesis into sources of data. Conduct exploratory data analysis.
•    ean: Creativity is the key. Build models, Integrate analytics into the IT workflow. Focus on building analytics with those elements of data that the customer demands, ensuring that the effort is proportionate to the value derived.
•    xtensible: Perfection by continually removing low-value data, poor-quality data and unnecessary data. Feed analytics insight back to the business process to refine the outcome on an on-going basis.

Case Study:
An FS Company SCALEs Using Data with History
From an analytics perspective, it is generally easier to work with data that has some history than it is to attack brand-new data sets. A large financial services organizations recently took this approach.
The company had billions of data records generated by the customer transactions on their platform. They wanted to create a program for the business users to help them systematically identify the nature of activities on their platform across the globe and thereby minimize the risk of conducting business in certain geographies. To achieve this goal, it combined just 3-4 data elements for a 3-month period from one of its real-time systems whose run rate was about 4 million transactions a day. The objective was to create ‘high definition’ portraits of its customer’s activities across the globe. This data set, combined with an interactive dashboard control, helped dissect data the right way; it enabled checks and balances to be put in place to control the risks associated with conducting business in select geographies.

This FS company
1.    utilized the SCALE methodology, with
2.    an appropriate use case that solves a business problem
3.    using data assets from existing consumer information for a shorter time
span of 3 months
4.    building the models and visualizations that support good decision making
5.    integrating the solution into the existing enterprise
6.    with an ability to extend the results to many more use cases for
minimizing risks.

The Result
The benefits they derived from this approach are three fold:
•    Ability to see and relate to the business benefits much quicker
•    Minimal upfront investments on tools, technology, and resources
•    Extensible to cater to the future demands of the organization
Consequently, the Big Data Analytics Program leadership received approval from the business team(s) to explore and deliver more use cases — this Big Data roadmap got them to their destination.