We are taking you to another website now.

Creating Exponential Value with Data Democratization

 

Kamlesh Mhashilkar
Head, Data and Analytics (DnA) Practice - TCS
22 September 2019

Data as a catalyst for exponential growth 

Data is a valuable resource and mankind has a knack of producing it abundantly. In recent years disruptive technologies have permeated our everyday lives and triggered ever greater levels of data generation. Unique ways of both creating and capturing data have emerged which provide numerous possibilities for optimization, targeting, and the generation of insights and innovation. But unless this data is available to, and used by, relevant people and programs, its value is dramatically curtailed. Democratizing data – ensuring it’s of, by and for the people, – is crucial to realizing its benefits. Its objective is to enable individuals and groups to use it at their convenience, where possible without the need for expert assistance. 

Data boosts our understanding, optimization, and productivity since it has reciprocal effects on the technology from which it breeds. Take, for example, Tesla, and its self-driving cars. A flood of data about road conditions, maneuvering techniques and driving best practices makes the technology possible. The capture and interpretation of data in this manner advances the technology and in turn, more data is generated. 

Digital-first companies with embedded Business 4.0 technologies such as Artificial Intelligence (AI), Big Data and analytics, Cloud, Internet-of-things (IoT), have an upper hand but legacy businesses can also benefit by democratizing data. Taking advantage of this data will help them thrive through faster decision-making, enterprise agility, improved customer experience (CX), operational efficiencies, and new products. 

Why do we need data democratization? 

Data has been called the new oil. Refining it into actionable intelligence unleashes its potential. This can be achieved at three levels. 

• Within organizations, it has the ability to fuel business projects across functions. 

• Outside of them, within the ecosystem, it can provide invaluable insight to and about the customer- and supplier networks, and other stakeholders. 

• When data is shared into a universe of academia, data scientists and policymakers, it acts as feedstock for new and innovative products and services. 

Democratization pushes organizations to rethink how they manage, distribute, and interpret data. Traditionally, data has been considered an IT asset, but this limits its value. When data, and knowledge, are made available to functions across an organization its latent potential is unlocked. New collaborations and innovation take root. Technology acts as a great enabler. With a balanced combination of machine and human interventions, the decision-making ability of an organization increases. 

Unleashing the true potential of data 

To democratize data, the effort needs to address three categories: Data literacy, data enablement and the creation of borderless ecosystems. 

Literacy: The foundation stone 

For data to be relevant requires a level of data literacy and an ability to understand its provenance, its value, its limitations, and risks. As the volume of data available increases, so do concerns around its governance and hence governance becomes a key facet of data literacy. 

At the heart of the data literacy debate are questions about where the data resides, who owns it, and how good it is. Digital technologies enable answers to these. But it is also crucial that leaders understand the implications of democratization for individuals within an organization: What’s in it for me? How will it shift the paradigm of decision-making? What does it do for career progression? How does it elevate the role of humans in an AI world? These questions are precursors to data democratization. 

Governing data democratization is essential to plan and hedge associated risks and avoid potential reputational damage. Consider the six dimensions in this checklist: 

  • Ethics: An ethical process requires understanding the purpose for which data is shared. Assessing whether the intent and basis for collection and sharing of data are published, and frequently monitored, is critical. 

  • Privacy: Adhering to privacy norms to protect personally identifiable information is key in maintaining stakeholder confidence. 

  • Trust: Globally businesses find themselves facing a trust deficit, in part because of the way data has been treated or mistreated. Issues related to trust with the entities in the data value chain, including regulators, must be identified and managed. 

  • Legal: A dedicated legal team focused on democratization activities should be appointed to oversee regulations and data related legal risks. 

  • Risk: Every business process involves data-related risks. An action plan to reduce its impact will strengthen the chances of success. 

  • Valuation: Data is a valuable, as well as expensive, asset. The democratization process should only be undertaken if it adds value and the effect on brand value, security, moral values, and principles can be monitored. 

Data Enablement: Capture, management, and consumption 

Digital technology makes capturing large volumes of information possible. This now includes IoT sensors embedded in products, images, videos, audios, and other paperless forms. Once the information is captured, the raw data has to be made usable which means converting it to a meaningful format. 

A global bank needed to have authorized data stores to meet regulatory compliance. Data used for decision-making in banking needs to be traceable and metadata tracks its origins, transformations, and approvals. Master datasets including customer, product, and channels information, need to be uniformly defined and managed across the organization. So, when regulators ask questions, it can now answer in 1-2 weeks, not 2-3 months. The bank’s stress-test ratings are up. 

Information on which decisions are based has to be trustworthy so its authorization is important. It cannot be done manually so digitization of the certification process in real-time, and its frequent audit, are pressing concerns. Controlling and authorizing ever-increasing volumes of real-time data is only possible digitally, whether it be structured, semi-structured, or unstructured. Configuration of data controls should be focused on activities such as the correlation of business-data elements, cross-verification for quality, cleanliness of masters, adherence to privacy and security norms, and records of the data lifecycle. This improves the trustworthiness of the data. 

Capture and management of information should be the responsibility of the business and not just of its IT function. This builds enterprise-level accountability. Data capture and management is a key enabler of democratization so that the right people in the organization can use it on their own instead of depending on help from experts. Until this phase is done properly, consumption cannot take place. 

Individuals prefer different ways of looking at data. We’re not all data scientists so a key aspect of democratization is the packaging of data in forms that encourage consumption. Three categories of services can be used: 

  • Data Services: These can be facilitated through mechanisms such as API, microservices, reports and dashboards, search engines and conversational agents. At the heart of these is the simplified semantic layer which is overlaid on top of different data stores. Such plug-and-play technologies encourage effortless consumption and business users find a value that can be extracted from data. These can also be extended to partners through controlled channels and new value streams can be sought from existing investments. 

  • Visual analytics: Most people understand visuals better than numbers. Illustrated data is easier to consume than a large content-heavy dataset. Visual depiction is gaining traction in the form of infographics and interactive reports. There may be a number of people in the enterprise who are good at data visualization and they should be enrolled. Self-service visualization tools that build appealing reports and help easily identify patterns are increasingly accepted in business user communities. Visual analytics is a key enabler of consumption. Augmented and virtual reality are gaining prominence with data scientists and business users. 

  • Conversational systems: Asking for the answer has never been easier. Conversational systems are changing our expectations of our interactions with machines. Whether it is a point data or a report or a document search a strong data services layout at the backend is required to fulfill such needs. 

People dependency may lead to knowledge leakage so establishing digital knowledgebases, as a part of decision management exercise, is wise. A combination of policies, procedures, and technology are needed to retain knowledge even if there is employee attrition. This knowledge, when democratized, can help more people to make better decisions. 

Beyond the Organization: Borderless ecosystems 

Making data available outside the organization creates opportunities for suppliers, partners and other members of the business ecosystem (including competitors) to create new value. Beyond the organization and its partners, regulators, policymakers, academics, and the general public may also benefit greatly from the broader availability of some kinds of data and even they can reciprocate the benefits. 

Opening organizational data to the ecosystem can lead to improved efficiencies, optimization of resources, and innovation. Insurance companies identifying fraudulent claims are limited by the data within their own organization. If this data were shared, the number of frauds could be reduced as the ecosystem is able to perform better vigilance and weed out offenders. Democratization here creates a win-win situation for the company, its partners and end customers. 

A global retailer opened up its data to suppliers (ecosystem) via an extranet. Suppliers can now analyze their performance and importantly compare it to that of others. They can plan better to meet SLA’s and strategize against the competition. All parties have access to the data which then informs negotiations and drives improvements for the retailer, its supply chain partners, and customers. 

A market research organization has developed new revenue streams by monetizing the data it and its ecosystem partners gather. Platforms are emerging as data marketplaces allowing partners to decide what and how they wish to share data commercially. 

Open data policies encourage the democratization of data. They provide individuals or businesses a means of accessing information to identify traits and risks and they can stimulate new business ideas. Open banking policy in Australia, and the UK’s PSD22, policy are great examples. Sharing information encourages the identification of customer behavior and risks. But organizations need to consider the latent risks data policies bring with them. Open banking policies may help to stop money laundering activities, frauds, and financial crime while opening sensitive information raises concerns about data privacy. 

Problems involving a large combination of data may prove to be overwhelming for a company’s limited set of data scientists. Opening the data to a larger set of people including academics, institutions, and regulators could be the solution. The drug discovery process in life sciences is a case in point. Research into new molecules requires possible combinations of multiple factors such as regions, demography, ethnicities, weather conditions and reactions to other molecules. Such scale and complexity create immense challenges – often too complex to be handled by scientists within a single business. Currently, the potential to discover personalized medication is enormous and deducing patterns for dosages is an important part of that process. It’s clear how life sciences can greatly benefit from the democratization of data at a universal level. 

Building a secure environment on the cloud and using containerization allows data sharing and experimentation without compromising security. Proper tokenization, masking, and pseudonymization techniques will further protect data whether it’s being shared within the organization, ecosystem or universally. A crowdsourcing platform with such capabilities can enable the global data scientists’ community to help organizations make better decisions. 

Dare to Democratize 

The governance checklist lets data-rich companies evaluate if and how they can increase their data’s value exponentially through democratization. If the outcome is positive, then they must develop an implementable strategy. Data security and privacy should remain in firm focus throughout. 

Education and skilling–improved literacy–contribute to the ecosystem-wide adoption of democratization. Breaking organizational siloes is a necessity to make data a mutual asset so that its benefits can be shared collectively. Without collective accountability and active support of all participants across the enterprise-ecosystem-universe, democratization cannot be accomplished. An enterprise-wide culture shift will be required if data democratization is to achieve its full potential. 

A global data revolution is underway. Leaders in their sectors will take advantage of their data to become optimized, innovative and agile. 

About the author(s)
Kamlesh Mhashilkar
Head, Data and Analytics (DnA) Practice - TCS

He heads the Data and Analytics Practice under Analytics and Insights Unit at Tata Consultancy Services Ltd. with the prime focus on strengthening solution delivery capability through integrated solutions and offerings (spanning BI, Information Management, Big Data, Business Analytics, and Artificial Intelligence), strategy and architectural excellence. Along with the thrust on bringing industry experience into analytical computing, his expertise span across Business Strategy, Practice Management, Corporate Research & Innovation, Consulting, Presales, and Coaching.

He has more than 25 years of experience in the field of Data and Analytics. He has been one of the founding members of BIPM Practice in TCS. He has been instrumental in accelerating the competency development and managing operations for bringing substantial revenue growth, analysts, partners, and institutional relationship and customized services under BIPM. He also led the growth in Big Data and Analytics under Digital Enterprise Services & Solutions Unit. He has conceptualized and built TCS frameworks and solutions in Business Intelligence and Data Warehousing branded as BIDS™, overall next-generation Data & Analytics Target Operating Model (DATOM™), Digitized Information Management Ecosystem (DIME©), Decision Fabric™ (Solution for Contextual Industry Offerings using different disciplines of AI components) and Data & Analytics Estate Modernization. He has 2 patent granted and 3 more submitted in the areas of AI and Information Management.

He has led many Data and Analytics solution deliveries in the domain of supply chain management, telecommunications, and financial services, which include Data and Analytics initiatives for Walmart, Bank of America, Reserve Bank of India, British Telecommunications and Sony Pictures.

He holds a Bachelor of Technology degree in Electrical Engineering from IIT Bombay.

×