Data Democratization – Moving towards faster, effective data-driven decision making

Annapurna Ramesh

Associate Data Scientist, Data Science COE, Central Functions, SG GSC, India

Sagar Anand

Product Owner, Emerging Technology, Innovation, SG GSC, India

“We are surrounded by data but starved for insights.”

-Jay Baer, Author, Entrepreneur, Marketing and Consumer Experience expert 

Data is powerful and valuable to businesses. It can generate actionable insights for organizations in a world of uncertainty. However, data must be processed, shaped, and analyzed before it can generate any useful insights. It requires for organizations to establish new sets of protocols and build a culture so that data is analyzed and used effectively in day-to-day business decisions. This research paper delves into various pain points of businesses striving to create a data-driven culture and suggests how ‘data democratization’ practices can help organizations achieve their data goals.

Gaps in current data practices

  • Data access

Traditional data practices have led to data dumps without proper access rights to suitable business users. As per a survey of business users, around 50% of business users need to wait for months to get data for analytics[1]. This hinders them from using the created data for analytics in a timely manner as the data often loses relevance.

  • Lack of data governance model

Data usage requires organizations to set up systems for data storage, quality check, and regular maintenance.  Today many organizations lack data governance models, resulting in an underutilization of data. An article published by McKinsey& Company says that only 1% of created data is finally used for analytics[2].

  • Lack of data skills

Business users need certain key skills such as data preparation and interpretation to be able to effectively use the results of data analysis. As per a report published by Deloitte[3], several jobs posted by

technology-driven organizations in 2020 show the need for analytic skills. The supply of talent in the market with these critical skills has not kept pace with the rise in demand for them.

Data Democratization – Helping make effective and timely data-driven decisions

Data democratisation can be defined as making data accessible to the average non-technical user for data-driven business decisions, without the involvement of IT. It aims to enable employees at all levels to comprehend data insights and make data driven decisions with ease.

Data-driven culture results in higher efficiency, responsibility, better work-from-home experience, faster decision making, better customer experience – all leading to improved business results and higher profits. In the past, it was difficult to create data democratization with traditional data systems and governance, however, it has become feasible now due to the advent and widespread usage of data cloud storage, data virtualization tools, plug and play data models, and the development of a data-driven mindset among business users.

Ways to implement Data Democratization

Data democratization requires multiple data layers depending on data policies and systems in the organization. The following diagram represents data layers for data democratization by Virtusa.

  • Data Mesh

Data mesh is a new approach based on a modern, distributed architecture for analytical data management. It enables end users to easily access and query data where it lives without first transporting it to a data lake or data warehouse.

  • Business Intelligence (BI) tools

Business intelligence (BI) tools provide real-time business insights. BI tools are connected to one or many data sources to present results in the form of reports, charts, and visuals. Benefits offered by BI tools include faster decision making, single point of truth and an increase in business transparency.

  • Cloud Computing

One of the ways to make data accessible for anyone from anywhere is to keep data in the cloud. It reduces data siloes and makes it easy for employees to get the data they want. Cloud computing makes massive computing power available to everyone. The flexibility of the cloud makes it a good fit for data analytics. By leveraging cloud platforms, petabytes of data can be digested from anywhere, on any device, by anyone, with answers understandable to all.

Building a data democratization framework for the organisation

Data democratization in essence is a set of practices that help create a data-driven culture and help employees to make data-driven decisions. This may involve multiple steps depending on the organisation structure and its businesses.

Organisations must ask certain questions before they embark on a data democratization journey. These questions are self-reflective in nature and can help organizations build their data democratization strategy. Systems and data ownership at Societe Generale Global Solution Centre (SG GSC) are explained through the answers to the questions listed below:

What is the need of deriving insights?

Timely and relevant insights which can be acted upon are crucial to the continuous improvement of a business. Businesses need to track key metrices such as profitability, revenue, operational expense, sales, attrition, etc. depending on their goals and stage of evolution. These metrices help organizations monitor their operations, take decisions based on the data, and perform better. It is for the organization to decide which metrices are valuable to their businesses and why do they need it.

Is the data available?

Data could be present across various systems and has different forms and platforms in which it operates. Metadata management can help in data discovery to understand if the required data is available or not. In SGGSC, Data Catalogue provides meta-data information such as data description, sensitivity, source, location, and ownership. These are available in the SG intranet.

Is data openly accessible?

Sensitive data should not be accessible without restrictions to anyone and everyone in an organization. There should be specific rules to govern data accessibility. Confidentiality, and Data Protection laws (ex: GDPR) are very critical in defining data accessibility rules. In SGGSC, data can be accessed through the SG markets APIs based on access provided by the data owner. During the entire data journey, no critical information should flow through the Chinese walls.

How to get the data?

In any organization, data roles and responsibilities are distributed among multiple participants. Hence it is very important to know whom to contact along with the procedure to get the data. In SGGSC, data manifesto defines the roles and responsibilities for data management. There is a distributed ownership ranging from data producer to data consumer.

Is the data of right quality?

Bad data is bad for business and data quality is everybody’s responsibility. If not heeded, it can damage client relationships, and generate breaks in the operational processing and poor decision making. In addition to this, poor data quality can lead to incurring fines and increased remediation costs for regulatory reporting. In SGGSC, we measure data quality based on 4 parameters – completeness, accuracy, timelines, and consistency. Average total percentage of these parameters should be >70% for a dataset to be considered of good quality. Any dataset below this percentage will require remediations.  

How do we achieve the required data literacy?

To implement data democratization successfully, employees are required to have certain data skillset. In SGGSC, data programs help achieve the required data literacy among employees. Data Citizen is a program which helps employees explore and understand data within their own perimeters. In this program, there are two levels – Data steward (study and analysis) and Data Citizen (development of ML model). After completion, mentorship programs that are in place helps industrialize the solutions in production.

Thomas George, Head of Data, GBIS, SG GSC, maintains that “Data is not anyone’s property but everyone’s responsibility. In this day and age, there is no time for IT to get involved to give access to data, clean the data, etc. these have all moved to “hygiene factors” now. Data Democratization ensures this through decentralised ownership and accountability and thereby establishes a sense of credibility to the owners. At SG GSC, this has been established with capable data owners, data quality managers and data stewards in some Business Units.

Data democratization is the collective responsibility of all employees and starts with a data-driven mindset demonstrated by the leaders in the organization. Leaders must help to cascade and encourage the practice of this mindset in organization structures, policies, and processes. The six questions mentioned in this paper shine a light on the steps and process of democratization. Data access, skill development programs and data responsibilities have to be an integral part of this process, so that employees do not face any obstacles in their data journey. Data democratization is an evolving concept and organizations need to stay updated with the new trends and technologies emerging in this field.


[1] Poll during webinar “Empowering Business users through Data Democratization” | 1Includes US values only. | Source: Mckinsey Global Institute analysis

[2] Straight talk about big data | Source: Mckinsey Digital data

[3] Tech looks to analytics skills to bolster its workforce. Author – Karthik Ramachandran (India)

Leave a Reply