The Role of the IT in the Analytically Driven Organization

Dr. Kenneth Elliott, Global Director of Analytics, Hewlett Packard Enterprise Services
69
97
23

Dr. Kenneth Elliott, Global Director of Analytics, Hewlett Packard Enterprise Services

Leading enterprises have teams of ‘data scientists’ sifting through terabytes of big data to find meaningful insights which can help their organization understand trends, recommend actions and predict outcomes. These organizations are on their way to master the growing deluge of information, find competitive insights in data, embed analytics into every business process and emerge as the market leaders in the information age. However, while this trend is growing, most organizations are not there yet. In order to emerge as an analytic leader, organizations need to cross the chasm from merely building analytic models to deploying those models into business processes and applications. This is where the role of IT in the analytic strategy becomes critical.

  ‚ÄčIn an analytically driven organization, IT expands its services and updates policies to embrace the agile nature of analytics   

IT must help bring analytics out of the shadows of the organization, deploy analytics into business processes, and provide an environment for exponential analytic value.

Bringing Analytics Out of the Shadows

If your organization is like most, analytics was first housed within IT as part of the ‘Business Intelligence and Data Management’ organization. However the agile and exploratory nature of analytics quickly became challenged by the methodical and design-first discipline of business intelligence and data warehousing. Traditional waterfall development methods were unable to meet the more fluid requirements of discovery-based analytics. 

As a result, any data scientists you now have are most likely sitting in one or more business units or in business operations—but, not in IT. On one hand, this provides speed and agility to your business. On the other hand, this limits the enterprise deployment and reuse of analytics across your organization. Such an arrangement suffers from redundancy of effort, escalating software and development costs, growing (hidden) support costs, data security risk, and a lower overall return on your analytic investment.

In an analytically driven organization, IT expands its services and updates policies to embrace the agile nature of analytics. This includes creating data discovery zones that provide the business with tools and the base data needed to enable data exploration and analytic modeling. This requires new service level agreements for things like data quality and performance for these business developed assets. This also requires a shift in IT focus from developing end-to-end solutions, to providing the tools and data to the business to support analytics. In the end, the business will get the agility they are looking for, while IT can manage costs, system performance and reduce data security risks.

Deploying Analytics into Business Processes

The role of IT becomes critical when the desire is to embed analytics into business processes, applications and machines. For business groups who do not collaborate with IT, the best they can do is build insightful models, gain intelligence and share results in the form of research reports or data extracts. However, if they wish to embed predictive analytics into call center applications, website, mobile devices, etc., IT is required to help build such solutions. Without a close relationship with IT this is either difficult to achieve or does not get done at all.

To address this, IT needs to partner with the business and bring its own team of data scientists to build analytic solutions. In most analytically mature organizations, data scientists sit within the business. This makes sense given the deep and ever-changing level of business knowledge required build effective analytic models. However, IT needs analytic experts as well to develop high performance business solutions for the enterprise. A good analogy for this duality is the relationship between concept car designers and automotive engineers. Concept car designers must be creative, innovative and unconfined in their design. Many ideas flop, but some are winners. These concepts are then handed to the automotive engineers who must translate the concept into technical designs which meet manufacturing feasibility, safety standards, legal guidelines, and cost limitations. In the world’s top auto manufacturers, this relationship between concept and production is well established and mutually beneficial. 

In the effort to build analytic solutions for the business, IT must bring their knowledge of analytics and work closely with the data scientists in the business to build winning solutions.

Environment for Exponential Analytic Value

Of course, the analytically driven organization is highly dependent upon technology. This is where IT is needed the most. Too many business groups are still building and hosting their own analytic solutions. The growth of cloud hosted options makes it even easier for the business to provision technology on their own. While these investments provide rapid results to the sponsoring business group, these silos weaken the overall intelligence an organization has about its business, lowers the value of the company intellectual capital and fosters incongruent customer experience. Not to mention the redundant costs, support and data security risks. 

It is up to IT to provide such a compelling set of technologies, with the benefits of the combined data and accelerators that the business has no need to go anywhere else.

Beyond the traditional business intelligence platforms, and the recently established big data platforms, the analytically driven organization needs extended capabilities to support the full analytic lifecycle from discovery to production. This includes:

• An analytic discovery environment to support flexible, fast, and short-term analytic research

• An open analytic workbench that enables data scientists to experiment with the best possible tools for the business problem

• Analytic model management capability that treats algorithms and business rules as company assets and moves them from laptops to secure central repositories for administration, monitoring, refreshing and legal compliance

• Decision and rules engines that enable the business to automate analytically based decisions within the appropriate business context and adjust business outcomes by managing rules

• Reusable micro-services capability that enable the organization to quickly leverage prior work and configure solutions based on components that are tested, effective and have a faster time to value

In summary, IT must not leave analytics to the business—it is a team sport. IT must provide the business with a safe and flexible discovery environment to foster innovation and exploration, provide a platform that enables rapid development of enterprise analytic solutions, and bring like-minded data scientists to collaborate to on solutions that can scale across the enterprise for the greatest return in analytic investment. Analytical modeling is just one component of what it takes to become and analytically driven organization. Only when IT and the business work together can they make analytics a competitive differentiator.

Read Also

Launch a Data Science Practice with These Five Questions

Seth Dobrin, VP & Chief Data Officer, IBM Analytics

Insights Are Only as Powerful as the Data Driving Them

Prasanna Dhore, Chief Data & Analytics Officer, Equifax [NYSE:EFX]

The 'Platform Mindset'- How enterprises can evolve data and analytics to deliver business value

Ritesh Ramesh, CTO, Global Data and Analytics, PwC and Oliver Halter, Principal, Information Strategy & Big Data, PwC

Industrial Data Analytics: An Edge to Cloud Strategy

Jay Allardyce, COO, GE [NYSE:GE]