Data Driven: An EA Perspective

In this article, I would like to share some thoughts on data driven strategy from the perspective of enterprise architecture. I hope it will help both IT and businesspeople see this subject in a different light and provide clarity in addressing prevailing business challenges in this space.

I am describing what should be obvious; however, sometimes it helps to restate the obvious just to level set the common understanding. An organized activity by a group of people who provide value-added products (goods and services) to customers for goodwill or profit is a simple definition of business. Human interactions and communication are intrinsic parts of business operations, and information must be exchanged among people to conduct various business activities to deliver the products.

Business cannot exist without information exchange. This exchange of information takes place directly between customers and customer-facing staff, and then a lot more information exchange takes place among employees from internal organizations providing necessary services to each other in support of the business operations. On any given day, millions of transactions are handled by staff. The amount of information exchanged obviously depends on the size and nature of the business.

If unlimited data storage capacity were available with negligible cost, then every organization could keep all the information exchanged in the form of historical records. They would store data pertaining to every transaction from the very beginning of its existence up to the current moment. It would be like keeping a digital copy of the entire history of the business transactions, capturing every state transition of the business, moment by moment.

With such a data repository, one can recreate a snapshot of the business at any moment in history and be able to inspect what the business was up to and how it performed doing whatever it was doing at that moment. Of course, in the real world, this is not possible yet; however, organizations can still come close to realizing that dream if they focused only on select few information elements pertaining to important and relevant transactions and only retained that information in the form of digital data.

Why should businesses care to keep such data around at all? Once the transaction is over, the data need not be kept around unless required by law, right? Well, that is the billion-dollar question, isn't it?

If business leaders realized the true worth and potential of the historical data at their disposal and understood what they can do with it, they will without a doubt keep that data and mine it to the maximum extent to advance their place in the markets. Those who fail to see the value of their data will be oblivious to what they might be missing.

The sky is the limit when it comes to harnessing the potential of data; it depends on how deep their pockets are. Data can be used to draw insights using various types of analytics. Every person in the data field by now knows the various types of analytics?-?

• Descriptive Analytics tells what happened in the past
• Diagnostic Analytics helps understand why something happened in the past
• Predictive Analytics predicts what is most likely to happen in the future
• Prescriptive Analytics recommends actions you can take to affect those outcomes.

But of course, there is no free lunch. Data management and analytics backed by AI is a costly endeavor. Ignorance about potential use cases for data analytics/AI and not knowing the cost-benefits of data and analytics capability within the business are the main reasons why businesspeople hesitate to enter the fray.

Enterprise data is the important and relevant data that I described above exchanged by internal and external users to conduct day-to-day business operations to provide services to their customers. When a business employs a strategic 'data-driven' approach, it means that the business has made a conscious choice of leveraging enterprise data to help make fact-based decisions at all levels of organization using data analysis and interpretation.

A data-driven approach enables businesses to examine and organize their data with the goal of improved service to their internal and external customers. By using data to drive actions, a business can contextualize and personalize its messaging to its prospective customers for better experience. They can monetize the data to provide new services.

Enterprise data consists of all sorts of operational and reference data involved with day-to-day transactions processed by hundreds of business systems. Organizations keep data about their employees, customers, products, various partners, and all sorts of other reference data specific to their industry. They keep data related to accounting and finance, sales and marketing, legal, policies, standards, principles, etc. ?and data pertaining to business planning and enterprise risk assessments. They also keep data about technologies used within the business and ongoing / planned programs. And so on.

Data can be structured i.e., it can be queried using database technologies and unstructured, i.e., all other forms of data such as files containing?-?text, emails, attachments, digital-audio & video, etc. including hardcopies and handwritten notes.

All this data is essential for business operations. Keeping this huge amount of data around for long can be expensive but it does provide additional benefits. Historically businesses have always had a good instinct about maintaining this historical data as it provided them a lot of useful insights about past, current, and future performances. Larger organizations used decision support systems for decades to gather and analyze large amounts of data synthesizing it to produce comprehensive audit reports.

With technological advancements, data storage has become affordable and accessible to all businesses, small and large, leveling the playing field. With the advent of AI now businesses can gain useful insights and make decisions for cases that they never even thought possible. AI can help humans make more informed decisions at a much quicker pace giving the necessary competitive advantage.

Imagine what lies ahead, with new large language and other forms of generative AI models being released every week, it is very possible for organizations use these inexpensive models and train them on their enterprise data and so using a natural language interface they can obtain all sorts of insights. This would be a powerful tool in the hands of their employees. Tech companies are already providing such AI services to improve productivity in various areas of business operations (e.g., Microsoft CoPilot).

Of course, proper care must be taken to ensure that the AI models are trained ethically and in unbiased manner so the generated content is not only accurate but also ensures ?models recognize the users and respond to the questions by applying right filters to only generate information that is accurate and allowed to be viewed only by the users with right access as per their job functions.

Enterprise data is only useful if it is shared by organization across all its business units anywhere on the planet. The data must be clean, current, and accurate and rendered in human consumable form. People should be able to easily discover and search for the needed information and be able to draw needed insights using AI tools without requiring any technical knowledge of the underlying technologies.

When data is not managed properly the value degrades very rapidly. Loss or theft of valuable data can result in significant financial losses for the business and their customers and so businesses must spend significant upfront time and efforts on data strategy planning and design i.e., effective data modeling, solutioning, security, and storage.

Not all business can afford to be data-driven or for that matter, they need to be fully data-driven given the nature of their business and the life stage of their business. For a business to choose to be data-driven, the leadership must fully buy into the idea. The data-driven approach is a strategic business decision. It demands large investments. Care and attention must be given by the leaders to ensure the strategic data initiatives are implemented successfully so it can start delivering intended returns.

The data-driven strategy fails due to a few main reasons.

Unclear strategic vision, goals, and objectives:

adoption of data-driven approach is a business strategy; it is not an IT or technology strategy. If this is not a top-level strategic business goal, then there is a very good chance the strategy will fail due to lack of sufficient investments.

No business leadership support:

even when the business decides to priorities and launch data-driven approach it may still fail if the leadership who are responsible for implementing this strategic initiative do not understand what they are getting themselves into. And due to their ignorance and reluctance to learn, the initiative does not get required the necessary executive attention and support and is more likely to fail. Key to success is education and creating awareness of 'the art of possible' with the enterprise data at the top leadership (including the board of directors) of the business.

Improper target operating model design:

Another key reason for failure of adoption of data-driven approach is the failure to design the right operating model for the data and analytics management practice. Business leaders must identify the right strategic problems (use cases) they must address with the new data and analytics management practice. Which will help identify the right services to be supported by the practice which then can help the operating model designers to ensure the right organization structure is designed with the right people make up with the required skills and ensure that the identified data services can be delivered using well-defined processes that are integrated well within the rest of the IT and business organizations and then finally right technologies/solutions are applied to support the data and analytics ecosystem for the enterprise.

Technology is just the means to an end; it must never be the focus of the data-driven approach.

IT led initiative driven by technology vendors:

Again, data-driven approach is the business strategy and not the IT strategy. If the formation of data and analytics management practice is initiated by IT especially guided by technology vendors (or for that matter, by IT consulting firm) then it is very likely to not get visibility across the business and due to lack of this visibility and recognition by business, it is likely to fade away as a result of unenthusiastic response from business (due to lack of business use cases that resonate with businesspeople) which very quickly dries out investments in the underdeveloped data analytics/data science practice. Most data strategy initiatives that are driven by technology vendors, end up adding to existing IT Operations cost usually because the businesspeople never really understood the rationale behind the technology solutions in the first place. It becomes another case of 'the hammer looking for nails'.

Just to summarize, adopting a data-driven approach in contemporary business operations highlights the hidden potential of enterprise data for informed decision-making and its indispensable role in facilitating business activities. Businesses should recognize the significance of historical data and transformative impact of advanced analytics and AI and harness the power of their data assets to maximum extent. Moreover, they should be cognizant of the challenges and opportunities associated with managing vast volumes of data and recognize the importance of strategic vision, leadership support, proper operating model design, and avoiding technology-driven initiatives. Businesses in essence should consider adopting a holistic approach to data management that aligns with business objectives and fosters a culture of data-driven decision-making to drive organizational success.

Author: Sunil Rananavare, IT Strategy Planning and Architecture (CIO Advisory)

Follow me to stay informed and share it if you found it useful.

?

The views in the article are author’s own and not necessarily of his employer.