Developing Successful Data Products at Regions Bank

Reading Time: 8 min 


AI in Action

This column series looks at the biggest data and analytics challenges facing modern companies and dives deep into successful use cases that can help other organizations accelerate their AI progress.
More in this series

Employing a product orientation has long been an important component of success in the software industry. Product managers shepherd new software from the earliest stages, where customer needs are identified, to the finished offering and beyond. As one description of software product management put it, successful product managers must balance the different perspectives of the technology, the business, and the user experience — all without having direct control of any of these domains.

Over the past decade or so, as software has increasingly incorporated data and analytics features, the idea of data products has become popular among digital native companies. These are software products whose primary purpose is to do something with data — collect, manage, analyze, or facilitate the consumption of it. Data products also typically involve some degree of analytics or AI models. Virtually every offering from Google, for example, qualifies as an AI-enabled data product.

Embedding data, analytics, and AI into products has been a game changer for people like Manav Misra, the chief data and analytics officer at Regions Bank. The financial institution has $161 billion in assets and is one of the nation’s largest full-service providers of consumer and commercial banking, wealth management, and mortgage products and services. Misra came to Regions with a background in software: He had been a professor of computer science and then worked in software and analytics businesses for many years. A product orientation became second nature to him. That’s a good thing, because when he arrived at Regions in 2018, he discovered both the need for a product focus and the opportunities it presented.

The Challenge: Data Science Isn’t Product-Oriented

One of the big problems with data science is that models simply don’t get implemented often enough. Data scientists like to create models that provide an excellent fit to their data, but that’s not enough to ensure their use in businesses. Deployment, which is when models are put into production, typically requires a number of other activities, including working with stakeholders, integrating a model with existing systems, changing the business process, upskilling users, and so forth. Multiple survey results suggest that model deployment rates are low — models end up being only experimental proofs of concept, or no one engages business stakeholders enough for the models to take hold. In a recent small poll of data scientists conducted by KDnuggets, the majority of respondents said that only between 0% and 20% of their models are deployed. Other surveys suggest that companies have had problems getting economic value from their AI investments — in large part because they don’t have enough of them in actual use.

One of the big problems with data science is that models simply don’t get implemented often enough.

When Misra arrived at Regions Bank, he found implementation levels and approaches among data scientists that were similar to those he had seen previously in his career. Scientists generally felt responsible primarily for model creation but not for any of the tasks involved in actual model deployment. “Enterprise software companies can’t hand something half-baked to an enterprise customer,” Misra said, “yet we were doing that in our analytics teams. We needed a major change in culture toward delivering complete solutions.”

A sense of collaboration between the data scientists and the people who deployed the models was missing. One group developed them and then handed them off to another for implementation. Few of the models were being used by the business.

Data Products and Data Product Partners at Regions Bank

Misra saw that the bank needed a change in direction, so he established the concept of data products and the new role of data product partner to manage them. He defines data product as “an end-to-end software solution to a problem that the business has. It leverages data and advanced analytics heavily in order to deliver that solution.” Data products may be intended for internal use within the bank or directly by customers.

Data product partners function as both data product managers and partners with the business to bring products to fruition and manage them. Each partner is assigned to a particular business or support unit of the bank. They need to be able to sit in on staff meetings with their business partners and understand that group’s priorities.

It wasn’t easy to find people for these roles. “The hardest thing was to make sure we had the right people,” Misra commented. “They had to have the mindset of both a business partner and a product manager and understand data, analytics, and AI.” Many came from the business side of the bank, though some also had analytics backgrounds.

The job of the data product partner includes understanding what the business unit and end user need, putting together development cycles starting with a minimum viable product, building complete solutions, and monitoring their use and effectiveness over time. During the early stages of the process, data product partners serve as bidirectional translators, connecting the opportunities of analytics and AI with business needs. At the later stages, they are responsible for assessing how the product is adopted and used, supervising how well the user interface is working, keeping track of how many people use it, and ensuring that the product delivers or exceeds the value promised in the original business case.

A New Process for Analytics and AI

Misra said that there are a few especially important components involved in the success of the data product partner role and the discipline of product management for analytics and AI initiatives. One is to ensure that the partner role is strategic, proactive, and focused on critical business needs, and not simply an on-demand service within the company. All data products should address a critical business priority for partners and, when deployed, should deliver substantial incremental value to the business. The teams that work on the products should employ agile methods and include data scientists, data managers, data visualization experts, user interface designers, and platform and infrastructure developers. Misra is a fan of software engineering disciplines — systematic techniques for the analysis, design, implementation, testing, and maintenance of software programs — and believes that they should be employed in data science and data products as well.

This product orientation also requires that there’s a big-picture focus, not just by the data product partners but by everyone on the product development teams. Misra says that even the data scientists now focus on helping to deliver a complete solution, and they are very pleased by the attention they get and the value they are delivering to the customer. All team members must always think about the end user and implement solutions that are engaging and likely to be adopted by them. They also must understand that a product requires continual monitoring and adaptation. Unlike a project, it isn’t a “one and done” effort. Resources need to be dedicated over the life of the product to ensure improved functionality and continued usability into future versions.

The data product partners measure the baseline performance before it is implemented and the impact of the data product afterward, including the results, revenue generation, and internal savings. They are also encouraged to promote their results to the business in an internal quarterly newsletter circulated across the bank, which also helps build awareness and drive demand for their partnership.

The Fruits of the Data Product Focus

Regions has been employing the product orientation for three and a half years. It has been quite successful, with the deployment of more than 10 revenue-generating/cost-saving products (with an incremental financial impact in the eight-figure range) and several more for internal support functions. For example, one product, called RCLIQ (Regions Client IQ), is a machine learning application providing relationship and contact planning for commercial bankers and client teams. It has 100% adoption among target users and provides early attrition alerts, monitors credit risk, estimates Regions’ share of wallet, prioritizes leads, and generates call planning insights. It has both increased revenue and lowered costs for the bank.

“You can point to a lot of aspects of RCLIQ that make it successful, but the No. 1 reason is that it is useful and additive for Regions’ clients, which in turn makes it valuable for our relationship managers and teams,” Misra said. “Like any good product development, you have to start with the customer and work backward. That’s where a data product manager can make a significant difference.”

A second tool, called Regions Voice of the Customer, uses natural language processing to enable omnichannel listening to customer feedback at key points in the customer journey. It integrates customer feedback with customer knowledge from another Regions data product. It is currently being used by several groups within the bank, including the digital banking and complaint response organizations, and is being expanded into an enterprisewide platform. It has saved over $1 million in annual payments to vendors and allows up to three times faster customer issue identification and five times faster resolution.

These successes have not gone unnoticed. At a recent senior leaders meeting for the bank, Misra’s group was invited to demonstrate a variety of successful data products. The occasion provided a forum to show the results of the group’s work and also sparked ideas and future use cases for the company’s senior executives.

Unlike many analytics and AI groups, Misra’s team can claim a perfect batting average of deployment of prioritized products. Ultimately, that success comes from the work of the team, as well as the prework devoted to building collaborative and inclusive business cases, recognition of the product partner role, and, most of all, the culture of deploying well-constructed and valuable analytics and AI solutions to business partners. In short, as Misra commented modestly, “The data product focus has really served us well.”


AI in Action

This column series looks at the biggest data and analytics challenges facing modern companies and dives deep into successful use cases that can help other organizations accelerate their AI progress.
More in this series

More Like This

Add a comment

You must to post a comment.

First time here? Sign up for a free account: Comment on articles and get access to many more articles.