As previously discussed, there’s an ever-increasing number of data sources and APIs available to help financial service providers embed innovation into their existing processes. However, combining all these data sources and APIs to create measurable business value is where the real complexity kicks-in. Our CEO, Alexis, explains the critical importance of getting your data layer right – and how our smart data models help you create your competitive advantage.
In 2020 alone, more than 600 new financial APIs were created, and the pace of innovation only continues to accelerate. Our integrated API marketplace allows financial service providers to access previously siloed data and integrate best-in-class point solutions more easily than ever. But “access” is simply the tip of the iceberg: the real challenge is creating measurable business value from these innovations.
The number of financial APIs has exploded in the last decade:
Consider a typical B2B customer onboarding process in a financial services context, such as international payments, credit and lending, insurance underwriting, or even more straightforward use-cases such as bank account opening. To onboard a customer, you will need to evaluate their application data against various “sources of truth”, and then offer a financial product as determined by your policy rules. Typically, this will involve orchestrating more than 10 different data sources for each operating market, each with their own schemas, consent flows, complexities, and economics.
Without the ability to easily share and understand the same data across these different sources, each integration will require a custom implementation, often resulting in costly data duplication, manual dependencies, and disjointed customer experiences. This is exacerbated by both the number of data sources typically involved, and by the complexity of the data they return.
Getting a consistent and complete view of the customer, and making sense of all the related data is increasingly challenging:
According to Experian and MIT Research, bad data and poor data management can amount to an astonishing 15-25% of revenues, which would put the cost to the financial services sector at more than $1 trillion. Unsurprisingly, financial services providers are expected to spend more than $120bn every year in the coming years on data management and risk analytics to tackle these problems – and that’s not even counting embedded finance players.
If integrating from scratch, engineering teams can quickly become overwhelmed by both the complexity of the data engineering tasks and the domain-specific business expertise required to make the data actionable to the wider company. Disconnection between data teams and business users is a well-known and documented issue. Under the pressure of launch deadlines, design compromises result in rigid and non-scalable processes. These become particularly apparent when the business grows to accommodate more users, new products, or international coverage.
Business users who consume the data such as risk operations teams currently work around these challenges by manually consulting each data source, cross-checking their contents, and assessing whether any discrepancies are significant. All this is very time consuming, error-prone, and hard to automate.
It’s no surprise that using tools like Excel to manually connect and reconcile different data sources is still the standard in many parts of financial services.
And these changes have a much broader business-wide impact. Financial service providers have to invest time and money on critically important non-core and non-differentiating activities, such as supplier management, data engineering, and constant maintenance – time and money that could be much more impactfully invested in optimising customer experiences or developing more competitive decisioning policies. As a consequence, it takes much longer to enter new markets or launch new products; and without a comprehensive view of the underlying data, managing and optimising the entire portfolio is almost impossible.
Most organisations have long recognised that having a data platform is not merely a nice-to-have, but a necessity. Value is unlocked through coherent and consistent data activation: generating actionable insights, improving customer experiences, and providing the scalable basis for process and decision automation.
In a financial services context specifically, data is used throughout the organisation and customer lifecycle – indeed, the “data layer” is arguably the single most important component of information system design:
Sikoia in the Financial Services Tech Stack
The ROI of getting the data layer right is measurable from Day 1:
The key role of a data layer is to abstract your data engineering complexity (such as data integration, maintenance, and harmonisation), allowing you to easily build scalable products and automate your processes.
We think there are five key capabilities that are required to power an effective data layer:
Connect – a data layer must be able to connect to all your key data sources, regardless of provider and source, and surface data from these in a consistent manner. We solve this with our integrated API marketplace, which has a library of turnkey integrations to the most important financial data sources and API providers. Our normalisation capabilities ensure that adding new data providers across core provider categories is straightforward, ensuring our own vendor-neutrality and allowing our clients to plug-and-play with their own suppliers.
Normalise – every data source, even those that are ostensibly very similar, will have its own nuances and complexities. A data layer must return normalised data to allow you to build standardised processes. Our proprietary Unified Data Models provide a shared language across these financial service API providers, powering a set of standardised, extensible schemas for each provider category (such as credit bureau data, company registry data, identity data, and so on).
Prioritise – different data sources may return overlapping – and possibly conflicting – data, and so your data layer must be able to prioritise data hierarchies, resolve data conflicts, and reduce error. In order to action API providers automatically, you may also need to “daisy-chain” across multiple sources to return the best data. For example, a company’s legal name may be different to its trading name; so it will need to be transformed to trigger adverse media or online presence searches.
Link – In any typical financial services process, you will want to evaluate multiple entities (company, person, assets) related to the same customer record. With a coherent view of all these entities, Sikoia’s data layer surfaces the relationships to optimise real-life application management. Most importantly, we support you moving beyond a single point-in-time snapshot of a customer record to a comprehensive view of the evolution of your customer profile as more data is collected.
Unify – achieving a unified customer view across API vendors seems impossible to many businesses, resulting in poor decision-making due to inaccurate customer insights and process inefficiencies. In fact, 89% of business have challenges creating a single customer view. Conversely, a well-designed and comprehensive data layer, with all the capabilities described above, can provide this view “by default”. As we are integrated into all relevant APIs and data sources, Sikoia is uniquely positioned to generate a unified customer profile and provide actionable insights over any individual provider used in isolation.
Improving your data quality is a huge business opportunity. Better data means fewer mistakes, lower costs, better decisions, and less friction in customer experiences.
But simply implementing a data layer is not the end of the story. We help our clients solve their specific business use cases and automate their key financial service processes. In our next blog post, we will highlight how we help our customers built modular and scalable workflows and best-in-class data-driven decisioning, automating key customer flows like onboarding, verification, and risk decisioning.