How to design a successful modern data analytics platform in an enterprise?
What are the two key characteristics that a successful data platform should have but that are often overlooked and ignored?
A modern data analytics platform, or big data analytics platform, or data platform, is an architecture and a working product that enables users to extract business value out of data, in the era of big data which is often measured by 4 Vs, veracity, volume, variety, and velocity. There are some prominent characteristics a data platform should have. Though it is trivial, it is worth repeating. It should act as a centralized repository, called a data lake often, that can store structured, semi-structured, and non-structured data in its rawest form. Its physical storage should be scalable enough to support the rapid growth of data across all the units in an enterprise in the era of IoT. The data security of it should be compliant with stricter local and global regulations like General Data Protection Regulation (GDPR).
Here are the 2 paramount but always neglected characteristics that can sail us to building a successful data platform.
1. User Experience and Design
It is the user experience that is important. The goals of a data platform are to empower users to apply analytics to solve problems and to make their life easier, and build a data-driven decision-making culture into the enterprise DNA. One key metric is that a successful data platform should enable employees to become data analysts to some level. This is what we called self-service analytics sometimes.
The success of it hinges on collaborative efforts from different teams who can put users at the center. Nevertheless and sadly, it is often seen that different teams (digital team, service team, project management team, development team, architecture team) start with ‘copy and paste’ technical specs from top-down or only focus too technically on their domain expertise instead of thinking from a broad-based view of where and what they can do to make a data platform more user-centric using their domain expertise.
These are internal walls. The Internal walls between physical infrastructure, service, digital, project management, architecture and design team need to be broken down. User-centric design is everyone’s responsibility. We need to think or rethink from a user perspective. We need to map their journey to a data platform. How can a data platform help them solve their pain points? How can it make their life easier?
2. Openness and Flexibility
A data platform needs to be open and flexible. Openness means that it needs to be accessible through public API, Java API, REST API, or Graphsql with proper data governance. Flexibility means you can leverage a data platform to achieve different tasks by using different languages, Python, R, or SQL.
The idea of openness and flexibility in different forms and types resonates with user experience and design because of a simple reason. We want to build a data platform that makes our users most productive regardless of their knowledge level or regardless if they are business users, intermediary data analysts, or trained data scientists. If the skill set required by a data platform is too stringent, no one will be able to use it no matter how much state-of-the-art technology stacks we implement in it, let alone seeing the value out of it from the C level. If the data platform is user-centric and user-friendly but not open in the sense that it cannot be integrated or cannot communicate with other systems in an enterprise, it will become an isolated and standalone application.
Enterprises that invest in these 2 characteristics boost their odds of building a successful data platform.
