Five Common Fintech Integration Challenges– and Best Practices for Addressing Each One (Challenge 2: Data Mapping)

Data mapping is a foundational integration process. For fintechs, getting it right allows them to onboard new customers in a reasonable timeframe while getting it wrong causes delays, frustration, and the risk of churn. In this second of a five-part blog series, I’ll share the data mapping insights and best practices that my colleagues and I developed from delivering numerous successful integration projects, helping fintechs connect to their customers' banking cores.?

Data Mapping Overview

Fintechs perform operations that exchange data with their clients’ banking cores and other systems. This requires discovery to identify the corresponding fields in each system, and then a process to map them to each other correctly. On many fintech integration projects data mapping is simply considered to be another integration task. Instead, data mapping should be approached holistically since it impacts fintech strategy and comes with several technical and human challenges.

Data Mapping Strategy

Data Mapping impacts a fintech’s go-to-market strategy and product features. For example, some fintechs standardize their APIs on a data model such as FDX and require this for all customer integrations. Other fintechs provide their customers with tools (for example, a no-code mapper) to enable them to map their banking core to the fintech’s data model. Rather than being an integration afterthought, data mapping must be included in a fintech’s strategic planning and product roadmap.?

Banking Core Operations?

Planning considerations must also account for how fintechs choose to map to their clients’ banking cores. Typically, there are three approaches, each of which has its pros and cons:

1. Map to a data model and require the customer to map their banking core to that model.
2. Map to a data mode and then build or license a connector from that model to the banking core.
3. Create pre-built mappings to most common cores.

Data Model Standardization

Most fintechs define and lock down their data model to provide a consistent set of operations and integration requirements (at least from the fintech side). Some do this by mapping to a standard such as FDX, BIAN, ISO20022, or CUFX. Each of these provides capabilities as well as gaps that fintechs need to address.

Custom Fields?

In addition, fintechs must account for and potentially map to any custom fields that their clients have built to extend their banking cores’ data models. This applies to the immediate (time-to-market) integration and long-term maintenance strategy.

Fintech Updates

As fintechs update their existing features or build new ones, they need to take mapping into consideration.

?

Technical Challenges

Data mapping has its share of technical challenges, many of which center around the difficulty of aligning two fundamentally different systems.?

Historical Differences

Integrating fintechs and banking cores brings one face-to-face with the generational difference between two systems. Most banking cores date back to the mainframe era and have a very flat data structure. Most modern fintechs were born in the object-oriented era and have data models with layers and hierarchies. As a result, mapping the fintech and banking core data models effectively requires bridging the two eras of computing.

Structural Differences?

Because most fintech data models support specific use cases, they tend to be small and well defined. Banking core data models, on the other hand, are sprawling in nature since they were designed to support comprehensive banking operations. Aligning these can sometimes seem like finding the right drop of water in a reservoir.

Updates and New Features

Because fintechs are constantly evolving with updates and new features, the challenge of mapping to the banking core is an ongoing one. In some cases, a new fintech feature may generate additional fields that have no corresponding ones in the banking core. When this occurs, fintechs and financial institutions must work together to decide where this information can be stored in the banking core. Mapping cannot take place until this process is completed.

Time and Effort

Fintechs and their customers are often frustrated by the amount of time and effort it takes to deploy a new feature or update to production. Frequently, the process for doing this is static and fragile. Any new field or new definition of a field requires refactoring the integration and this includes developer effort, regression testing and redeployment.

Human Challenges

Data mapping also comes with human challenges, most of which pertain to setting expectations and leveraging available SMEs

Integration Context

During the sales cycle, fintechs familiarize customers with the features and the value proposition that their products offer. Integration is often treated as an afterthought that confuses customers when they are subsequently asked to provide banking core access, documentation, field definitions, and SMEs who can walk fintechs through these things.?

Customer Expectations

Data mapping is a collaborative effort between fintechs and their financial services customers since fintechs require customers to provide detailed information about the banking core data model and to validate test results. This can be problematic in cases where customers expect a turnkey solution that requires no effort or resource commitment on their part.?

SME Availability

Data mapping requires fintechs to obtain detailed information from their customers’ banking core SMEs who can walk them through the vagaries of their banking core data model including custom fields. In many cases, however, the financial institution is unwilling to space the SME’s time. Sometimes the financial institution has no banking core SME on staff which means fintechs must gain an understanding of the banking core data model through trial and error.

Data Mapping Best Practices

Data Mapping is relatively painless when fintechs set customer expectations, proactively capture mapping information, and position themselves to manage future changes. The team that addresses fintech data mapping with a trial-and-error approach will struggle, while the team that does so with a well-defined process is likely to succeed. The key to the following best practices is to implement them early, typically at the final stages of the Sales cycle.?

Set Expectations

1. The best way to set customer expectations is to provide early education, (usually a brief presentation) about what data mapping is and the level of collaboration it requires including SME availability. ?
2. Providing visual detail is very effective. I have worked with fintech leaders who hosted training sessions in which they walked customers through specific application features and explained how each of these had to map to data in the banking core.
3. Train the Sales team on the role of integration and how they can help set customer expectations during the sales cycle.?

Perform Detailed Discovery

1. Prior to the project kickoff meeting, prepare and send mapping templates and relevant questions to the customer. These are things you can walk the customer through in that meeting.
2. Work with the customer’s banking core SMEs to fill out the mapping templates. Depending on the customer’s organizational processes and the SMEs, this could be a smooth process, or it could be challenging. Be patient.
3. Engage SMEs to create the definition-of-done and to validate results during the testing phase.

Plan for Management and Maintenance

1. Document the process for updating existing features or rolling out new ones. Review this with the customer for input and signoff.
2. Create a RACI matrix or similar document and include customer SMEs in this process.
3. Provide early notification of new features or maintenance windows and treat each like a discrete project.

Conclusion

Depending on how it is handled, data mapping can be one of the most challenging fintech integration tasks or it can go smoothly. To be successful, fintechs must give data mapping a strategic focus by including it in the product roadmap and go-to-market approach. Success also requires a tactical focus such as providing customers with early education and arming them with the tools and assistance needed to surface important details. In next week’s article I’ll cover Challenge 3: Middleware Selection.