News & Analysis

Data Democratization: The key to Unlocking Data-Driven Organizations

data science

By Vinod Ganesan

Today, data makes the world go round and with newer technologies being explored there is need for organizations to use data to the fullest to stay ahead of the curve. Studies have shown that data-centric organizations make better strategic decisions, have higher operational efficiency, improved customer satisfaction, and generate more robust profits. In fact, Forrester predicts that such organizations are on track to make US$1.8 trillion annually by 2021.

One might add that it is a lot easier to talk about becoming data-driven as an organization in theory but in practice, there are a number of hurdles to overcome – starting with sourcing and onboarding the right kind of data professionals that are currently quite hard to find. LinkedIn reported that the demand for data scientists in Singapore grew by 17 times from 2013 to 2017. To make things worse, another study found that only three in 10 C-suites and directors in the republic were confident in their ability to interpret, understand, and work with data.

Democratize that data

Limited number of data scientists available can pose a serious challenge in the organization’s efforts to become data driven. However, this need not be a problem if organizations take the approach of making data available, readable, and understandable to all concerned employees within the enterprise. By democratizing the use of data, you place the power of leveraging valuable insights within the reach of all teams and help them make informed decisions a lot faster, thereby unearthing several opportunities for the business without the intervention of a specialist.

Organizations have come to realize that there is no need to centralize the task of extracting insights from data to a specialist. Especially with the increase in the amount of data churned out daily, it’s constructive and beneficial to empower more employees with diverse expertise to make data decisions and improve business agility and performance.

Airbnb, for instance, through its Data University, strives to educate every level of the organization in basic data science. With a 100-strong Data Science team working on experimentation, data analysis, and machine learning, Airbnb has built data tools and a stable and scalable data infrastructure to empower all its staff and data scientists. Following specific training sessions customised to the focus and work of each team, employees can comprehend and visualise how to leverage data in their day-to-day tasks. This helps decrease the number of ad hoc requests that data scientists receive by 50 percent, allowing teams at large to rely on data for strategic decisions and to measure their results.

Low code to the rescue

Some enterprises might be hesitant to have their non-technical workforce interpret data as they might be concerned that those employees may not have the skills to read and derive the right set of insights from the data they read. Here’s how you can tackle that roadblock –

  • Self-service analytics: Deploy data analytics that have features like natural language query, visual data discovery, which are easy to use. Non-technical users can undertake complicated analytics with no external help, with these devices at hand.
  • Go low-code orno-code: Traditional computer programming can be complicated, there are newer tools that allow business users to build machine learning models using user-friendly features such as graphical user interfaces and drag-and-drop modules. Now that business users don’t need the help of a technical expert; software delivery is sped up significantly.The enterprise’s sales force, for instance, could leverage a no-code platform to build a machine learning-based tool that recommends relevant products to upsell and cross-sell, based on a customer’s purchase history.
  • AI/ML: As always, use technology to the fullest to gain maximum benefits for your business. AI/ ML must be industrialised so that it can ubiquitously operatein the background of systems and applications. Enterprises can then build, deploy, and scale AI/ML applications in an easily repeatable, predictable, and automated manner. When you have systems that can constantly assist non-technical, business users discover newer insights, they are encouraged to drive continuous transformation across the business.

The underlying and most important as aspect for organizations trying to become more data-driven is a solid foundation that allows effective data management. A solution here is the implementation of an enterprise data cloud using which organizations can:

  • Manage connected data workflows across multiple cloud and on-premise environments
  • Support multiple analytic functions – from streaming and big data ingest to the Internet of Things and machine learning
  • Maintain strict enterprise data privacy, governance and security across all environments
  • Maintain control of their future by ensuring zero vendor lock-in and maximum interoperability

Even if organizations are deploying low/no code solutions, they should ensure that their employees have basic data science and analytical skills. UOB, for instance, runs Open Learning sessions twice a month in addition to internal training. These are very popular withemployees as participants can clarify any doubts with UOB’s data analytics experts. While exchanginglearnings across functions, employees realise how to adopt data analytics in their own tasks. At the same time, UOB can identify areas where the organization as a whole may benefit from additional skills development.Not only do data literate employees take more ownership and responsibility of their decisions when driven by data, but data literate organizations also average a three to five percent higher enterprise value as revealed by the Global Data Literacy Index by Qlik in 2018.

Data can only help those organizations that use it optimally and effectively. Once analysed and insights actioned, it can help take businesses to greater heights of success. Coupled with the fact that data scientistsmight not be abundantly available anytime soon, Indian organizations need to transform their operating model to truly unlocktheir data potential. Democratizing the use of data is the only way to weave it into the organizational culture thereby continuing to reap business benefits over the long term.

(The author is country manager – India, Cloudera, and the views expressed in this article are his own)

Leave a Response