It is often stated that 'Data is the new oil', suggesting that data, like oil, holds immense value but remains unusable if unrefined. While this comparison simplifies, it effectively highlights two significant trends:

• A fierce competition to capture as much customer data as possible, waged not only by Big Tech and social media giants but also by major e-commerce players and supermarkets through loyalty programs.

• Growing investments in deriving insights from this accumulated raw data. The advancements in AI play a critical role, yet they merely scratch the surface. Beneath, substantial investments are required in tools to gather, cleanse, structure, and store data. The complexity and associated costs can be substantial.

The financial sector, rich in valuable data, has historically underutilized this asset for several reasons:

• Strict regulatory oversight ensures data usage within the financial sector is heavily controlled.

• The sector’s reliance on 'trust' is paramount; a mere hint of distrust can threaten a bank’s survival. Thus, banks are cautious about jeopardizing trust by misusing confidential financial data.

• Challenges in data capture and processing due to the siloed organizational structure and reliance on outdated mainframe systems. These systems, while robust and reliable, complicate the extraction and analysis of necessary data.

However, the landscape is changing. More banks are establishing data lakes and similar warehouse solutions to centralize data for analytics, alert management, and AI model training.

Among the most insightful data types is the day-to-day payment information of customers — either retail, SME or corporate. This data offers a detailed view of a customer’s activities, showing when, where, and what the customer purchases and receives. Analyzed over time, this data enables highly accurate customer profiling.

But also, this data remains highly underutilized, not only because of the above-mentioned reasons, but also because of certain specific challenges linked to this data, i.e.

• Payment data is often dispersed across various systems, with different payment systems employed across bank branches and customer segments (e.g. retail versus corporate customers). Additionally, also linked to the type of payment, i.e. a domestic payment takes a different flow than an international payment, an internal transfer, card payment or a payment linked to a bank commission or bank transaction (like term deposit or securities).

• The sheer volume of data demands specific technology (like distributed NoSQL technology) for efficient, near real-time analysis.

• Insights often require complex modelling to correlate multiple payments over time (time-series correlation).

• Payment data must be enriched with both internal and external data sources, like fraud detection, sanctions screening, and AML tools, or details about the counterparty (such as name, sector, activity type…​) to support all possible value-added services.

Fortunately, the data quality is high, thanks to standardizations such as SEPA in Europe, though some normalization and cleanup are still necessary.

By systematically collecting payment data, financial services companies can unlock a multitude of use cases, gaining a competitive edge through:

• Sales insights, such as 'Next Best Offer,' 'Churn Prediction' (e.g. ongoing outflows to accounts at competitors on same counterparty name) or 'Cross-selling Prediction' (e.g. high inflows of money).

• Value-added services based on payment data analysis, such as payment reconciliation with invoices, linking payments with additional information (like guarantees, manuals, ticket info…​) or providing personalized financial management (PFM/BFM) advice, like giving insights in income & expenses and providing recommendations how to reduce/optimise expenses, generation of cash-backs, calculation of ecological footprint (i.e. Carbon calculator)…​

• Improved, real-time integrations with customers and third parties, enabling detailed transaction information access beyond the minimum required by regulations like PSD2, allow (corporate) customers to easily retrieve via API their payment data, with even possibility of a push (via e.g. a webhook) to inform customers near real-time of received incoming payments, but also allow partners like PFM/BFM apps or accounting platforms to retrieve up-to-date, high quality and near real-time payment data.
  This would allow also to enable more cost-effective (compared to card payments) Account-to-Account (A2A) payments initiated via external APIs, or allow bank to tailor their outreach by better segmenting customers based on anonymized payment data. (cfr. JPMorgan Chase has launched a digital media business enabling advertisers to target its 80 million customers using transaction data).

The potential for creating value from raw payment data is enormous, provided the data is centralized consistently and reliably. Specialized tools are highly recommended for this purpose.