Analytics for Business Success – Effective Strategies & Indian Success Stories

The Tata Sons’ company employed advanced analytics comprising a range of technologies, such as machine learning, natural language processing (NLP), and semantic analysis to address the issue. The hybrid retail giant was able to develop a deep understanding of site visitors’ preferences by studying their social profiles, brand affinities, recent activities, and household and macro-economic data about them with the help of analytics.

Step into the future

From finding ways to increase business efficiency, to trimming costs, retaining high-value customers, determining credit risk, fighting cybercrime, identifying new revenue opportunities, and preventing fraud, advanced analytics can transform businesses across every vertical. With Indian organisations increasingly being data-driven, analytics has matured from being a good-to-have to a must-have technology.

And once organisations get analytics right, they can focus on the next step – artificial intelligence (AI). Implemented well, a powerful analytics-driven strategy can help Indian organisations compete better, differentiate effectively, innovate faster, and deliver better business outcomes.

Analytics – a must-have

In 2017, the total analytics/data science/big data market in India was estimated at $2.03 billion in revenues, growing annually at 23.8%, according to a study by Analytics India Magazine and AnalytixLabs. Research firm Gartner observed that Indian organisations are rapidly migrating from conventional reporting to sophisticated analytics tools, by adopting an architecture-oriented approach. The outlook for AI, which is built on a strong foundation based on analytics, is highly positive as well. A report by the International Data Corporation (IDC) and Intel India predicts that over 70% of Indian companies will embrace AI before 2020.

The analytics path to value

There are some key imperatives that CIOs should consider when it comes to delivering positive business outcomes driven by powerful analytics:

• Prepare data centre infrastructure for analytics –

  To successfully and effectively leverage analytics, organisations must get three infrastructure pieces in place—performance, security, and storage. IT infrastructure must support diverse analytics workloads, including real-time big data deployments (such as with Hadoop), in-memory databases (such as SAP* Hana, or SAS* in-memory analytics or Oracle* Exadata), a big data processing framework (such as Apache Spark*), or streaming analytics (such as Azure Storm or Apache Flink).

• Consider in-memory analytics -

  Traditional data warehouse models are limited by single systems and software to accommodate both transaction processing (OLTP) and analytics (OLAP) on the same hardware, using the same row-based data structures. As a result, data is replicated and reformatted into multiple schemas and aggregates to support analytics. A real-time data hub model, by contrast, uses an in-memory database to store data in a combination of row and column-based data stores while keeping all of it in-memory. This minimises the need for multiple copies of data in multiple formats. Near real-time analytics can be performed directly on operational or transaction data. The in-memory data hub can incorporate unstructured data using scale-out NO-SQL data stores, such as Hadoop, which have a growing number of in-memory processing capabilities. SAP Hana and Oracle Exadata are some leading in-memory solutions.

Harness enterprise data effectively

This involves laying down strict data governance and security policies, unlocking data silos to make data accessible, and making it usable at scale via data modelling, cleansing, normalisation and transformation.

Nearly 52% of the 606 global corporate IT decision makers surveyed consider “improving existing IT capabilities to promote agility and innovation” as the key driver of their firms’ IT transformation. Organisations must increase scalability, interoperability, and interconnectivity of infrastructure systems to be able to support increased data volumes and advanced analytics.

Universal applicability of analytics

Analytics is relevant in nearly every sector. From agriculture, BFSI, healthcare and manufacturing to telecom and beyond. Yes Bank*, for instance, uses analytics to design targeted campaigns and cashback offers for its customers. Automating the monitoring-to-settlement process, their solution credits cashback amounts to customers’ Yes Bank savings accounts or Yes Pay wallets.

India's leading private sector power generation and distribution company, Reliance* Power, employs analytics for condition monitoring to detect and avoid critical equipment failures. For Reliance, equipment failure not only meant maintenance hassles but also huge losses—running into crores of rupees—and severe reputation damage due to service disruption. The utility deployed a condition monitoring and diagnostic system, a predictive analytics solution, helping its managers take informed maintenance decisions quickly. As disruptions were reduced, so were losses, resulting in improved customer satisfaction levels.

Reference architecture

While the exact components of infrastructure architecture for analytics will depend on specific data types and workloads, here’s a reference model that businesses may find useful:

• In-memory analytics such as from SAS or SAP HANA
• Open source tools such as SPARK, Hadoop, etc.
• Cloud-based streaming and data management architectures, including no-ops models
• Security at the perimeter, access, visibility, and data levels -- for example, employing Apache Sentry and Cloudera* Navigator for Apache Hadoop
• Optimised fabrics for data movement between data lakes or warehouses and compute infrastructure, such as the Intel^® Omni-Path Architecture
• Infrastructure foundation comprising Intel^® OptaneTM solid state storage, NIC, software tools and libraries and Intel^® Xeon^® Scalable processors

Analytics and AI require a strong foundation

A well-defined data management strategy coupled with the right infrastructure can help organisations accelerate their digital transformation journey smoothly and securely.

A critical aspect of accomplishing digital transformation through analytics and AI is a strong infrastructure foundation that’s suitable for these advanced technologies and workloads. For instance, existing Intel^® Xeon^® Scalable processor-based infrastructure is already capable and optimised to help accelerate analytics and AI goals.

Intel^® Xeon^® Scalable processors form the foundation for modern analytics infrastructures. By combining scalability, agility and performance, Intel^® Xeon^® Scalable processors give organisations the flexibility to build on top of analytics and add machine learning, cognitive computing and deep learning capabilities to their workload processing tasks, making AI a reality.

Further, SSDs from Intel^® and Intel^® OptaneTM Persistent Memory help enterprises achieve scale and efficiency. Intel^® OptaneTM is a non-volatile technology that provides a combination of high throughput, low latency, high quality of service, and high endurance. Coupled with the hardware, software libraries and frameworks like Intel^® Math Kernel Library and Data Analytics Acceleration library make advanced analytics seamless for customers.