In the financial services industry, Morningstar—most often recognized as an investment research and insight company—is a bit unique. And that uniqueness stems from the fact that it seems to be uniquely everywhere.
In its 37 years of existence, Morningstar has grown to touch seemingly every aspect of the financial services ecosystem. It’s trusted for its independent data and research, which helps empower investors to make informed investment decisions. But it also serves a broad spectrum of financial industry professionals: asset owners and managers, financial advisors, and issuers. Overall, its versatility makes it one of the most impactful organizations in the financial world.
But by 2016, the sprawl that helped turned Morningstar into an industry colossus was forcing the company headlong into an existential challenge. Better said, two distinct but related challenges—data unification and, downstream of that, data utility. Yes, Morningstar’s insights have long been an industry leader. But the data that informed them was siloed, its storage system was aging, and the company knew it needed to quickly set up back-end infrastructure to best serve investors. To keep pace with the consumption demands of all its various audiences, Morningstar had to do something—fast.
These challenges aren’t particularly unique; they’re emblematic of those currently faced by countless enterprise organizations. What compounded Morningstar’s complexity was sheer data volume. For the past 17 years, Morningstar has stored every single piece of activity that has happened across 260 markets—not just executed trades, but bids and asks and plenty of other context to help make sense of demand. And that was just a piece of its overall puzzle. “Daunting” feels reductive.
So when James Rhodes, Morningstar’s chief technology officer, arrived five years ago, he knew what he was up against. He also knew he needed a strong technology provider to help tackle it: Amazon Web Services (AWS).
“This was where we could leverage providers like AWS on how we design something for the capacity and needs we have today,” Rhodes said, “that we know will scale in an efficient and effective manner for what we're predicting will happen in the future.”
Trillions of needles, thousands of haystacks
Frankly, Morningstar had a lot of data—its largest data set (coming from its market business) is roughly six petabytes, for example. That works out to about 6,000 terabytes; for context, even a single terabyte could store about a quarter million photos taken by the highest quality iPhone setting. To further complicate things, all of that data was stored across thousands of separate databases. Five years ago, Morningstar had approximately 4,000 employees—and held its information on about 9,000 on-premise databases, few of which could seamlessly communicate.
“I used to joke with everybody that we gave every employee two databases when they joined the company,” Rhodes said.
In many ways, AWS—long an industry leader in cloud and data solutions—was built for the challenge. By providing the consultation and infrastructure Morningstar was desperately seeking, AWS started moving the organization to a secure, consolidated data lake that could begin stitching together reams of disparate data. By roughly 2019, that work was paying off: Morningstar had created a new enterprise data platform (EDP) strategy. It was finally set up to tackle its second challenge—data utility.
Even working from a more consolidated data platform, Morningstar was still utilizing traditional production processes and manual handoffs between teams. That reality could often stretch production timelines for insight reports, for example, across nearly an entire year—by which point the market would likely have changed. In financial services, research departments can’t risk outdated information. To evolve, Morningstar had to craft a business built for the future.
“Like with financial markets, some things are time sensitive,” Rhodes said. “If it takes you nine months to get a time sensitive thing out there, you’ve missed the opportunity and the value to the market is not real.”
Rhodes calls that his “aha” moment; once again, AWS was a prime solution. Its vast data, analytics, and machine learning services were built to not only help customers store enormous amounts of data, but also help them process and use that data. By implementing a variety of AWS solutions—and rethinking how Morningstar took insights to market—Rhodes’ team saw an opportunity to completely transform its operation.
A key step in that process was the creation of Morningstar’s Analytics Lab. It would help streamline data access to help its quantitative research analysts (quants) and data and engineering teams quickly and efficiently dial up analytics and metrics to produce reports for their clients. The lab also created a specialized user interface to democratize the information, helping others make sense of it.
Analytics Lab is comprised of four elements, all of which lean heavily on AWS products: a data lake on Amazon S3; a core analytics platform that incorporates Amazon EKS (Elastic Kubernetes Service) and ECS (Elastic Container Service); a UI portion powered by Amazon EMR and JupyterHub; and a capability Morningstar calls Notebooks that exists within Morningstar Direct (the company’s investment analysis platform). All told, the Analytics Lab platform could provide a far more modern experience for two distinct audiences: quant research analysts, and end investors through the financial professionals that use Notebooks. On the one hand, users in quant research and data science roles would be able to create a Notebook exploring a specific topic—say, a comparison of mutual funds—with other members of the team seamlessly embedding rich HTML graphics and widgets to help illustrate the findings. The exhaustive process of compiling information, documenting assumptions for investor transparency, and passing it all along to development teams—something that used to make months—could now be done within a unified interface in just hours: not days, not weeks, and certainly not a year.
“What you’re left with is basically a notebook that just looks like an interactive, rich web experience,” Rhodes says. “The broad user actually has no idea that there is any code.”
See how Morningstar is closing the “last mile” in how it produces analysis for investors.
All the while, Analytics Lab’s second unique audience—the end investors that Morningstar is so fiercely dedicated to serve—get access to new insights that were previously unavailable, helping them make more informed investment decisions. While investors don’t use the Notebooks tool directly, they can now consume the content financial professionals produce through the tool far more effectively. If today’s investor diet is predicated on real-time market insights, Notebooks was suddenly Morningstar’s secret sauce.
Real-world impact
An example of why this ability to be nimble is so important: toward the end of 2020, then-President Donald Trump issued an executive order blocking American investors from investing in some Chinese companies that, in the administration’s mind, had too close of a relationship with the Chinese government and its military.
Morningstar began hearing from clients immediately. A rush of investors wanted to know whether their holdings contained such securities. Notebooks was born in that moment. Morningstar data and quantitative research teams quickly created reports identifying mutual funds and other investing products that would have exposed investors to the suddenly-banned Chinese securities—a process that usually would have taken months, meaning investors would have also spent months with an exposure that could have resulted in penalties. But with Analytics Lab, data scientists could suddenly scan upwards of seven terabytes of data for key information in just 10 to 15 seconds—and act accordingly.
The end result was astounding. “We developed (the reports) in a day and pushed (them) out to our users,” Rhodes says.
Hear more from James Rhodes on how Morningstar used Notebooks to help assess banned securities.
All told, Morningstar has produced more than 500 different Notebooks—and plans to produce far more in the months to come. With its EDP serving as a strong data foundation and Analytics Lab powering a new way of communicating its research, Morningstar has more than just transformed its operational reality—it’s pioneering an entirely new way of empowering investors across the world.
“I learned a long time ago that interesting doesn’t mean valuable—if we ended at interesting … that doesn’t translate to value,” Rhodes said. “The best way we empower investor success is by providing transparency and insights into very murky space. This platform and end-to-end lifecycle allows us to provide that transparency when it matters most.”
This story was produced by WIRED Brand Lab for AWS.
