Data lakes and central data teams were created to facilitate insight-driven decisions, yet in many cases they have become bottlenecks choking the flow of insights to the business. The demands of maintaining the data and its repositories leave most data teams little bandwidth to respond to the business’s data requests. More importantly, banks’ data is often siloed by line of business, by function and by channel.
This not only makes data management more complex and costly, it also causes redundancy and risk and inhibits sharing and collaboration. The first step to overcoming these problems in 2023 is a shift in mindset: data needs to be regarded as the oxygen that fuels the bank’s every action, rather than its CO2 that accumulates as an inevitable by-product of its actions.
The value to the bank and its customers should be recognized and it should be treated the same as a product. There should be product owners whose job is to identify and define the use cases, clean the data, and set up the APIs and structures that allow different teams and systems from all parts of the bank to access it. These product owners should promote the data and help the business to use it as effectively as possible. It’s not going too far to say they should be incentivized to help the bank meet data-usage targets. This is where the data mesh comes into its own. A data mesh is not a new tech product—it’s a change in mindset, approach and structure. It has the potential to be, to data, what agile was to our ways of working.
It starts with a business use case, which might be the identification of fraud or commercial loan underwriting. Based on the use case, a data product owner assembles the relevant data from all its disparate locations into an organized product, making it easily accessible by means of APIs. Just as any other bank product owner does, the data product owner then works with the business users— data scientists business analysts, operations, external partners and more—to help realize the business value. JPMorgan Chase created a data mesh to improve its fraud detection. It appointed a data product owner who pulled data from debit and credit card usage, check spending and other sources into a data product that allowed the bank to reduce fraud costs without compromising data governance.20
In a true data mesh the data is all connected, even if it is located in silos across the bank. What it does is create a marketplace for producers and consumers that democratizes data and allows users to analyze cross-domain data on their own, rather than putting their analysis request into the queue for the central data team to process. This in turn helps them create business value while driving down costs. By providing a single view of all data it helps eliminate redundancies, simplifies data governance and promotes re-useability. In 2023 we will be hearing a lot more about this new approach. Although data mesh has become a buzzword whose definition is murky, the power of data as a product is clear.
The real opportunity will be in finding the right data products to radically change business outcomes. Consider commercial banking. Today, commercial banks collect vast amounts of data from each customer—income statements, balance sheets, legal information, ownership data, ESG reports, to name a few—and store it in different places all over the organization. Whenever the customer applies for a new product, from a different division, the bank asks for this information all over again. The real opportunity, for banks that organize their data as a product, is to collect the information once and then use it repeatedly.
New commercial loans could be offered in hours, rather than days or weeks. Customer insights could be unlocked to proactively sell new products and AI could help relationship managers in real time to dramatically improve customer interactions. Data mesh, together with data product owners, is not the solution to all data problems. Yet it could allow banks to unlock the value trapped in monolithic data lakes and data silos—and to fundamentally change the way banking is done. Data, when treated as a product, has the potential to transform the foundations of banking. Banks don’t lack data, they lack the means to unlock it, and simply and consistently turn it into actions. Having a data product manager and mindset could be the key to unlocking the value trapped in
Veritass have worked with many data scientists to explore opportunities present in ‘big data’ harvested from client repositories. Whether interrogation or extracting new insights we are here to help your teams get the most out of a new type of banking gold.