FEATURES Spring 2018

Big Data discovery hosted in the cloud is your gateway to the Next!

Jeff Silverman, Grant Thornton, LLP

The Evolving Status Quo

Have you heard the horror story of a radical new business development that brought a company to a standstill? Whether it is tainted food products that necessitate a global recall, an unstable telecommunications device that can explode or even unexpected fiscal results due to unforeseen labor costs, the use of business analytics (BI) to resolve these issues can become desperately important – even faster than you can say “new analytics approach!”

The rapidly evolving business landscape requires swifter methods to redefine problem statements for the business and sometimes delves into questions that have never been asked. How can business leaders handle a new onslaught of questions if their analytics tools are only oriented to answer the common, day-to-day questions?

Discovery Toolsets as an Entry Point into Insight

Information discovery toolsets are vastly different than traditional business analytics. Information discovery tools can access unstructured data, moving business intelligence from a static and structured report-driven process to a fast and flexible user-driven process. An information discovery centered approach loads existing data in its current state as the first step to drive reporting requirements, eliminating the need to spend time creating a technical data model (or deal with the multiple changes to that model throughout the process). See Figure 1.

Figure 1

Information discovery tools navigate both structured and unstructured data, allow for every piece of data to be loaded and will iteratively review what pieces add value to the desired analytic. The reporting dashboards at the front end of a discovery tool allow for rapid prototyping of the solution, choosing the data elements and metrics of value without spending weeks creating the “perfect” report that invariably misses something.

The information discovery approach to solve analytics problems takes days to weeks instead of the usual months that a full software development lifecycle can demand for traditional data modeling of business analytics. Once a new reporting question is defined and answered (solving the immediate issue for a company), those insights can be applied to traditional BI to allow for scaled reporting and repeat review. Essentially, information discovery shapes the answers, while BI can industrialize the solution once the issue is stabilized.

Accessibility Via the Cloud

How can a traditional business intelligence user use an information discovery tool if it isn’t already in their technology stack?

When that challening circumstance occurs, the last thing a company has time for is an installation of unfamiliar technology upon on-premises servers that probably are insufficient to handle the new requirements. The answer lies within the cloud!

Infrastructure as a Service (IaaS)
Information discovery toolsets such as Oracle Big Data Discovery (BDD) or Oracle Endeca are resource hogs requiring robust platforms to process. Furthermore, some data sets that are truly unstructured can require additional processing for Hadoop tools such as HDFS, HIVE, SPARK or many of the other exotically named Big Data applications.

Using a hosted cloud environment as infrastructure can eliminate the need for pricey on-premises hardware and can allow a more measured approach to utilize a discovery toolset to solve an immediate problem. Implementation of such a method can be installed in two to three weeks, or faster with an implementation partner that already has the IaaS up and running with the appropriate tools. After initial setup, data can be fed swiftly to garner instant insights on those critical new business questions that move much faster than waiting for an on-premises solution.

Platform as a Service (PaaS)
Oracle is finalizing the incorporation of information discovery toolsets (accessing Big Data) into their Oracle Analytics Cloud (OAC), a hosted cloud platform that already houses BI and Essbase multidimensional cubes. The juxtaposition of information discovery and traditional BI allows for quickly addressing new questions with Big Data that can then be repeated and fortified for the enterprise with standard reports in BI/Essbase.

The PaaS method makes accessing new information even simpler than utilizing an IaaS model, as the software is already integrated from day one. Using an information discovery approach may feel unfamiliar at first for traditional BI developers. However, the immediate access to the tool along with proper coaching from a practiced partner will allow for swift modeling and access to interrogate the data to find answers to those new questions that can then be modelled for sustained reporting in the cloud.

Figure 2

Test Drive Information Discovery Via the Cloud

  • Ease and affordability of the cloud combined with the insights garnered from an information discovery approach provide swift answers at low investment costs.
  • Systematic use of Oracle Big Data Discovery and Endeca provides the ability to ingest structured and unstructured datasets.
  • Use of IaaS or PaaS methodologies accelerate the adoption cycle to use the toolset and can allow a “try-before-you-buy” approach.

Test Drive Information Discovery Via the Cloud

Wary executives anxious about the costs of an unproven tool can relax; the ease and affordability of the cloud, combined with the insights garnered from an information discovery approach, will provide swift answers at a low investment cost. The systematic use of Big Data tools provides the ability to ingest both structured and unstructured datasets and acts as a bridge into the enterprise dark data that holds the answer to the current crisis.

For example, if a product has issues, the discovery toolset can look at commentary logs on complaints; if there has been spoilage of certain food shipments, the toolset can look at disparate logistics logs for patterns. Moreover, such tools can graphically display those data points to help shape an immediate pattern recognition and provide a template for follow-on, standard BI to emulate. See Figure 2.

Utilization of IaaS or PaaS methodology accelerates the toolset’s adoption cycle and can allow a “try-before-you-buy” approach to answering the questions at hand.

Conclusion

It can be a vicious cycle. Corporations have many questions that need to be answered, yet are loathe to invest in a new product without further information. Tools such as Oracle Analytics Cloud can break the cycle with minimal initial investment, quickly sift disparate data sources, frame new analytic questions and experiment with previously unanswered subject areas.

Jeff Silverman is a senior manager with Grant Thornton, LLP and is a leader within their Business Analytics practice. Certified in in over 20 different technologies, Jeff specializes in assessment and planning of analytic solutions.