apacciooutlook logo

Centricity of Data Science in the IT World

By Harpreet Kaintel, CIO, ZenithOptimedia Group

Centricity of Data Science in the IT World

Harpreet Kaintel, CIO, ZenithOptimedia Group

Over the last decade, data has emerged from the vaults (or servers) to occupy the space of a key organizational asset. This has brought the incumbencies of an asset upon data, with a primary responsibility of creating specific value for the owner organization.

This has triggered a shift in the existential question for IT; from ‘how to connect and store’ to ‘value it is adding to the business.’ IT function has been gradually evolving from being storage and safety provider to connecting the organization, internally and externally. But today’s data order threatens this legacy built around the traditional systems and solutions.

The massive scale of this shift requires IT to move from evolution to revolution and drive a crucial link between data infrastructure and business application; essentially upgrading the function from a reporting system to an intelligent decision system. This is the domain of data science, which has the wherewithal to turn data into an asset and a competitive advantage.

A few structural factors of data flow layout the basis for this fundamental shift:

  • Dynamic nature of inflows: The days of research and reports have been supplemented by big data, predictive modelling, real-time optimization etc. The near-time speed of incoming information needs to be matched by agility to turn it into actionable intelligence. This calls for a fresh approach to technology and knowledge assets.
  • Multiple nodes: The biggest challenge is the variety and complexity of data available. Compared to a few researched or monitored variables earlier, today there are multiple nodes of data available for each problem. This is where the observational approach of graphing and charting doesn’t suffice anymore. It calls for capability to build meaningful links and interplay between different nodes, and this is the essential play of data science.

These changes are redefining the fundamental approach and key focus of data solutions in multiple ways, a few highlighted below:

  1. From ‘how’ to ‘what’ we store: Content is the new king and with growing sources and variety, the question is increasingly what does business need and how fast. IT is increasingly called upon to be an active collaborator, not only in the process of discovery and acquisition, but also to bring abilities of linking these pieces to the decision systems in a dynamic fashion.
  2. From ‘storehouses’ to ‘solutions’: IT is required to increasingly move from managing data dumps to providing active solutions which support decision making. The rise of virtual storehouses in form of cloud is bringing about significant efficiencies and creating opportunities to focus on solutions which have a direct impact on business.
  3. From ‘reporting’ to ‘decisions’: While pushing the right information to the business desks in a timely manner is still an important component, but creating intelligence is becoming an equally essential part of delivery. The amount of variables at play and ability of data sciences to link and interpret these algorithmically is the value creation system in decision space.
  4. ‘Timeliness’ to ‘real-time’: Another big change is the importance of real-time, which is an important upgrade to the timeliness concept. The important implication of this is the need for speed and integration of decision systems into technology delivery systems.
  5. ‘Visualization’ to ‘analytics’: The real value of data needs to be unleashed in two tiers, one by making it part of organizational culture and two, through bringing in specialized science. These two tiers can be loosely defined as: ‘data in work’ and ‘data @ work’.  Former relates to quick access and visualization with some simple analysis and needs to be embedded into everyday business processes. On the other hand, the harder part of data@work, a domain of specialists needs to be the core delivery of the data and analytics teams. This part should have its focus at creating business specific solutions and then embedding and industrializing them.

The enormous speed of change is leaving no room for a gradual transformational approach. What is needed is a tiered surgical approach to build almost a fresh eco-system. Here are a few practical principles which will guide these developments moving forward:

  • Adaptability: IT has been very adept at replicating and outsourcing models to leverage the scale of uniform technology solutions across organizations and industries. But with data becoming an owner specific asset, it needs customized solutions for every organization. This calls for nimble and adaptable approaches, instead of large commitments and creation of heavy legacy systems.
  • Business integration: IT needs to become part of organizational delivery system and stop behaving as a supplementary service. This needs IT to build business understanding and actively integrate with business teams to maximize value of data. The concept of distributed data teams across business groups is the way forward. On the other hand, it is equally incumbent on business functions to increase their appreciation of data and analytics so that they can collaborate effectively with the specialist data teams.
  • Business accountability: Finally, IT needs to get off the expense account and move to being a part of value creation. It needs to step up and make itself accountable to businesses. Data science is its bridge as it helps define the ROI of all organizational actions and holds the key to link data to business performance.

This is a crucial but transformational time for IT, either it earns that slot on the table or face relegation to the back-office. The function needs to collectively evolve from supplier mindset to being an integral part of the organization and make a concerted effort to change the definition of I from ‘information to intelligence’.