The Trellance Data Blog

Giving Credit Unions a Sustainable Competitive Advantage with an Industry Data Lake

Posted by Alex Beversdorf on Oct 22, 2018 1:03:00 PM


Tips to Secure Your Data Lake is the first of four episodes for the Data Lake BIGcast Series. John Best, the CEO of Best Innovation Group, brings in Rojin Nair, General Manager of Fintech Solutions at Celero to discuss data lakes and how a collaborative credit union data lake could revolutionize the industry. Celero is a well-established Canadian fintech company that provides a wide variety of services to the banking industry. By managing financial transaction processing and offering leading technology solutions, they successfully maintain over 80 credit union banking systems.

Emerging Technologies

We live in a fast-paced and continuously changing world. The steps we’ve made in the past decade with technology have been remarkable and there are no signs of this growth slowing down. Predictive analytics, artificial intelligence, and machine learning have just begun to crack the surface of their full potential. That’s why understanding industry trends is so important, so we can get a better grip on what the future might bring.

Existing fintechs are ready and chomping at the bit for this new era of digital transformation, while new ones are quickly forming to capitalize on this great opportunity. The future world of business will be data-driven, with an increasingly complex technological structure. That is why fintechs like Celero are crucial as they can help their clients (credit unions) better optimize and manage their technology capabilities. Rojin expands:

The challenge for organizations is that there are too many choices with technology. With 4,000 Fintechs out there, it’s hard to find the one that fits best. They all look good on the surface, but when you start to talk about integration with CUFX (the standard language for credit unions) they run into trouble. We serve as guides to our clients, to try and help them with their digital journey.

Diving Into Data Lakes: Get Your Ebook

Celero’s Data Platform Journey:

In Tips to Secure Your Data Lake, Rojin explains his experience at Celero as an Advisor about two years ago. He traveled around meeting with many credit unions and discovered one distinct trend that went unnoticed. That was the sudden influx of interest to start leveraging data to move beyond simple reporting. Overall, executives were very eager to try new things and explore the untapped market. These indicators led Celero directly to OnApproach, Rojin says:

For us, we found that the best way to move forward was to not look at the technology, but to look at a layer that was standardized. So that is when we found OnApproach. They were the only one who had a patent and followed the CUFX schema (one common language between credit union data). We were then able to take their platform and deliver the first fully-integrated analytics platform to our first client in exactly six months time…Throughout my 25 years in this integration initiative, that time frame is very hard to come by. Normally it takes six months alone to understand the data mapping. But, in this, case everything from mapping, to integration, to the five predictive models that we created, only took a total of six months.

OnApproach is a Credit Union Service Organization (CUSO) that enables credit unions to leverage the true potential of their data through an innovative data platform, M360 Enterprise. This is a prime example of how OnApproach’s software and Caspian Data Lake can benefit more than just individual credit unions. The CUSO model is designed to capitalize on the power of collaboration, giving vendors, fintechs and credit unions mutually beneficial opportunity in a data-driven ecosystem of information.

Privacy and Security Regulations

Sharing data across hundreds of credit unions obviously raises questions about security. Credit unions have access to very personalized and important data about their members, and it’s their responsibility for keeping it secure. Data encryption, along with other measures of security, make protecting member specific data easy. How is that possible? Rojin explains:

With any predictive model you’re not looking at anything like a member ID or member name. It is more so behavioral science modeling (i.e. what an 18 years old’s behavior looks like). The focus isn’t on their identity but rather on their persona. Focused more so on age rather than who that individual actually is.

Rojin continues to talk about governance and what that would look like from a credit union perspective. There are two basic ways to go about governing the data being fed into a data lake. The first one gives the control to the individual credit union and makes them the judge of the valid inputs that keep the confidentiality of their members. And the second one is switched to put the trust on the owner of the data lake with certain control parameters that need to be met. He goes on, “Both options are available, and you have to draw your conclusions based upon your risk value proposition and strategic alignment to see what the best fit for you is.”

What a Data Lake Brings to the Table

Generally, Rojin talks about there being two types of data, structured and unstructured data. “70-80% of the total data for any business fits into this unstructured format. At Celero, we have always been good at looking at the other 20% of structured data and surfacing some value from that.”

A data lake can make collecting unstructured data easier for its users as there is more data at their disposal. This is achieved by implementing one big data pool. Data lakes are very beneficial as Rojin explains, “They are cheap, very fast and a beneficial economical value proposition for any business model. It helps in identifying and using more of this unstructured data. Now suddenly, you’re liberating 100% of the value from structured and unstructured data combined.”

Data Lakes are a very simple example of how economies of scale can benefit businesses with regard to the use of data. The more data there is, the better. It makes predictive models much more accurate and useful as there are thousands of data points to build off of. In closing, Rojin believes that there is “no magic here, it’s a scientific method… If you can automize, aggregate and standardize all of this data together, then proportionally the quality of your predictions in any aspect of analytics will improve greatly.”

Listen to the podcast:

Stay up to date - Subscribe to our educational blog

Topics: Collaboration, Podcast, Data Lake