Data Discoverability, Consistency & Re-Usability: The key to Business Agility and Operational Efficiency

Akshey Gupta, Chief Data Officer, Bandhan Bank | Sunday, 26 March 2023, 12:17 IST

Aspiration to continually improve, and generate new avenues of business & automation are some of the aspects chased by every enterprise to have an edge over its competitors.  Personalization as a concept is not just limited to customers but it is expected from data to auto-nudge business teams for opportunities and daily tasks. It should have the capability to determine the priority, based on dollar value, impact on people by volume, and also soft factors like empathy and public sentiment. Governments and New regulations require more transparency and lineage of Decisions. The ultimate objective is Trust and Confidence. For a financial organization, this may relate to financial stability and in turn safety of the public money invested; similarly, for an FMCG or Pharmaceutical, it may relate to trust in the quality of the product delivered and its traceability to underlying constituents.

Any organization spread across multiple geographical locations may easily have 20-30K reports spread across MIS, throughput, regulatory, compliance, accounts, etc. Additionally, there may be close to 200-300 analytics, data science, and AI use cases. This is fuelled by multiple applications generating data at an exponential rate, combined with social media.

Robust Data Engineering and Governance processes are key to churning data at an astronomical rate and keeping it consistent. Handshake implementation, Dependency management, and API integration between multiple processes be it file exchange or system handover of data are very important. Though this sounds simple, it is still a missing component in sizeable enterprises. As an example, when the succeeding process starts at a pre-scheduled time, without completion of the preceding process, resultant reports/data sets come out blank. Dependency has dimensions of time and data to be managed.

A monitoring view of data crossing different stages, like a train traveling its journey via various stations, is required to check the failures and create actionable events. This not only reduces manual handovers but has the capability to create 30% capacity savings.

In a quest to better understand markets, and consumers, new methods are applied to study past data trends and build new use cases in space of Analytics & AI. Enterprise data modelled across Dimensions and Facts being updated consistently, should create a base that can be reused for any new requirements. It may be noted, Data preparation and cleaning constitute 60-70% of the overall time building the Data Science use cases. This will help the team save its resources and effort, thus would place the organization ahead of the building curve.

An Enterprise Business Matrix should be generated and published, listing business journeys across products, with its related Dimension and Fact objects. This enables Businesses to Discover & Self-serve. Using a Single Source of Truth adds to the effectiveness of the model, and consistency in Regulatory reporting reduces the model building time by 40%, gives sizeable capacity savings in the range of 25-35%, keeps a check on stale data sets, and removes redundant processes occupying a sizeable Compute and Storage.

Data Quality is a long-standing concern across domains. It causes an increase in Operational expenses toward customer service calls, degradation of customer confidence, and regulatory fines. Incomplete addresses, Mismatched City-State relationships, and inconsistent parameters like Aadhaar, PAN, etc are some examples. Exception reports based on Data Quality rules need to be defined & shared regularly with Business and Application owners. This can benefit Financial Industry significantly in the digital onboarding of customers and their KYC and Re-KYCs. The overall objective is to improve the Quality scorecard with tangible outputs.  

A cohesive approach, eliminating multiple versions and sources spread across departmental Data sets, is a fundamental foundation and approach towards a ‘Data Driven Enterprise’.

It is a journey with a vision to continually improve & grow!