Enhancing Critical Data for A Major Australian Financial Institution

The Problem

In order to provide quality wealth management suggestions to clients, this bank’s wealth managers and advisors rely on technology to collect massive amounts of customer banking data and crunch large numbers to draw portfolio performance insights and trends. Mantel Group identified the following in its key examinations:

• The ageing technologies that traditionally powered their wealth management have begun to show difficulty keeping up with the massive surge in the amount of data that has been generated over time and continues to accelerate at an exponential rate as the days go by. Of late, the reliability of legacy systems has dropped so low that rerunning failed batch jobs have begun to overshoot daily SLAs.
• These problems have led wealth management business applications to gradually become slower over time. The freshness of critical banking data necessary for wealth management advisory is dropping at a fast rate. Lack of fresh data availability from disparate systems in a fast, event-driven way is increasingly turning into a major bottleneck, preventing wealth advisors from being able to provide a customised and timely wealth management advisory to their clients. Of late, data freshness has dropped so low that data can often be up to a day old, and the lack of an up-to-date daily transactions ledger has restricted wealth management advisors to a snapshot of data that can be up to 2 days old.
• Clients are missing out on great opportunities to make use of wealth management products and services in order to time their investment moves to attain profitable positions quickly. The inability to provide customised and timely wealth management services to clients is affecting overall customer experience as well as advisor satisfaction, thereby leading to a drop in customer conversion and retention.

The Approach

This institution needs to quickly solve the problems of interoperability between disparate banking systems and data standardisation. They also need to elevate data freshness by leveraging the power of streaming real-time data and event-driven processing. These critical enhancements are necessary in order to enable wealth advisors to obtain data updates within seconds of changing, in order to provide accurate and timely wealth management advice to their clients in an ever-volatile and fast-paced investment banking environment.

The following key challenges were identified:

• Banking solutions built with the cutting-edge technologies of yesteryear are suffering from capacity and performance bottlenecks today. Bulky, monolithic software running on-premises on virtual machines and dedicated hardware is getting harder and more expensive to upgrade and maintain at peak performance. It is getting harder to find, hire and retain technical resources that can maintain critical banking software built with legacy technologies, which is a major risk. The inability to scale ageing banking systems is taking a toll on customer experience.
• Over time, various banking departments have developed custom solutions to handle their own data in bespoke ways, thereby creating inconsistencies, duplication and barriers to interoperability between systems. Non-alignment with modern open-sourced and global data conventions and standards is holding back the innovation and modernisation of systems. This inability to modernise is adversely affecting the quality and accuracy of wealth management advice that advisors can give their clients.
• The decline in data freshness has been attributed to the numerous steps and hoops of a growing number of slow, intermittent batch processes that have been historically put as tactical stop-gap solutions over the past few years. The inability to acquire and process data within seconds of creation/change is reducing the freshness and timeliness of critical insights available at the advisor’s disposal to make profitable wealth management advice in a timely manner.

The Solution

This bank set out to reimagine and rebuild an engine for customer insights on the Google Cloud platform to power the future of banking and wealth management.

To accomplish this, bulk and streaming banking big data siloed within dozens of disjointed applications will be securely ingested and processed at scale and in real-time. The insightful business intelligence metrics that the online portal for advisors and other consumers need to empower wealth managers and clients, will be modelled, computed and unlocked within seconds of a change in data.

Fully managed (serverless) products and services on Google Cloud will be leveraged to scale out/in per demand/volume of data. Elastic autoscaling will be used to optimise operational costs, thereby eliminating capital expenditure otherwise spent on the maintenance of physical hardware used in on-premises data centres.

• Working closely with stakeholders, Mantel Group identified major bottlenecks and opportunities to simplify complex business processes. Complicated workflows that had become convoluted over years of repetitive tactical fixes were streamlined. This allowed us to gain access to critical banking data raw (facts) as soon as data was created/updated. We successfully standardised disparate data into uniform, globally renowned data formats and conventions for increased interoperability.
• Working closely with Google, Mantel Group architected fully automated big data processing ETL pipelines to ingest batch and streaming data securely from several sources in real-time. We wrangled and modelled the data at scale to aggregate and standardise data from a number of incompatible data silos. By leveraging powerful big data processing products on GCP such as PubSub and Cloud Spanner, we unlocked critical business insights with the optimal schema and performance.
• We securely integrated an engine for customer insights with an online web application to make critical banking business intelligence intuitively available to wealth managers, advisors and customers through an online portal for advisors.

Key Products/Services Used

• Cloud Spanner
• Streaming Dataflow
• Firestore
• Docker Containers
• Google Cloud Storage
• Google Kubernetes Engine
• Cloud Build
• Google Observability Suite
• Pub/Sub
• Cloud Run
• Cloud Functions
• BigQuery