Mohammed Rupawalla, Vice President and CTO, Digital & Data, Mphasis, outlines the five ways in which data can become trapped, and what steps organisations should take to avoid these situations
During the last five to seven years, the idea that data is the new oil has gained currency. Beyond the obvious benefits to both corporations and society at large, there is a deeper analogy where the raw product is "trapped" and must first be extracted, cleansed, and refined before its value can be unlocked. The need to extract the trapped raw data and convert it into a refined product for commercial use has caught the attention of business leaders and technologists alike.
There are five main ways in which data can become trapped. Let's examine each one and discuss how to overcome it.
1. System or Application Trap
Traditionally, business data is generated in one of three ways. First, as a business buys and sells raw materials, finished products and services, it records transactions with various systems and applications. Second, corporations collect additional data to understand the health and direction of the company and the behaviour of customers and markets. And the third way - data generated by a company can be further supplemented with information purchased from external sources.
Data liberation concerns the first kind. A typical mid-sized corporation has a few hundred operational systems that capture business data. These systems report on current status and can advise on tactical business decisions. There are two obvious limitations - doing intensive analysis in the operational system can impact its responsiveness, plus all the analysis is limited to one system. Building a data platform is the clear answer here, as data has a longer half-life than processes. However, as organisations have been building data platforms for the last two decades, they must deal with a considerable legacy landscape before new systems can be implemented. It's necessary to visualise existing systems, choose what to modernise, and manage dependencies while reducing technical debt.
2. System or Application Owner Trap
We would ideally design the data platform to handle large amounts of data and analysis, but application owners are fiercely protective of their data. Some concerns are legitimate, surrounding compliance, confidentiality, and unauthorised access. However, there are also irrational fears around data sharing, loss of control, and value extraction. Businesses have the responsibility of avoiding this second trap and stepping up as the owner of corporate data.
3. Organisational and Cultural Stagnation Trap
The organisational boundaries within a company form a deep chasm - individual lines of business guard their data quite steadfastly, except when an enterprise-level initiative commandeers the data for a company-wide purpose. Overcoming this requires embracing a new mindset. Business leaders have to drive data liberation principles to remove unnecessary barriers to data sharing within an organisation. Before new systems and data sharing practices can take hold, leadership has to drive a culture change punctuated by success stories from the business. Technology can then step up to build these principles into the data platform.
4. Data Trap
The significance of investing in data quality and enrichment cannot be understated. Data sets are constantly expanding and evolving, making data the fourth trap. It is easy to be overwhelmed by the sheer volume of the data set, and striving for perfection can keep you from reaching your goals. The good news is that modern data and ETL platforms provide a rich set of quality management tools out-of-the-box.
5. Workflow Trap
Disruptive business models turn the status quo on its head. They require reimagining the world, sharpening key competencies and creating new differentiations. Data and AI/ML-driven systems can facilitate disruption by allowing for easy experimentation. However, the application landscape remains stuck in fixed linear workflows, forcing business thinking to flow along pre-defined, constricted paths. This is the fifth and final trap.
Established user journeys are largely a one-track linear sequence. We need cognitive journeys that take a user from one track to another based on user or external data, aided by relevant cognition. While the Netflix or Amazon type of recommendation model has been widely adopted, we are still in the early stages of applying AI and ML to guide more complex workflows.
Data exchange amongst enterprises is growing exponentially, and leaders now have to think beyond their own four walls about the broader ecosystem of consumers, suppliers, and competitors. By taking steps to avoid these five traps we can realise the benefits of a data-driven strategy. These steps must be implemented from the top down - leadership mindsets have to change in order to model and drive the culture changes that will support the new, liberated system.
This article was produced in association with Mphasis