Goldman Sachs Takes Aim at Interoperability, Analytics with Marquee Enhancements

When Anne Marie Darling started working at Goldman Sachs 21 years ago, trading platforms were largely closed-off systems. But with the advent of cloud, greater adoption of APIs, and the embrace of open-source tools, the investment bank—and, for that matter, the capital markets as a whole—is increasingly opening up one of its key platforms to the developer community in a move that could help the bank develop new managed services for users.

Goldman’s new ethos can be seen in the enhancements being made to its Marquee platform, which provides institutional investors with market views, hedging tools, and trade execution across multiple asset classes. The firm is in the process of creating an ecosystem that is more interoperable with other Goldman applications so that a user can get everything they need around information, risk, and execution in one place, rather than having to bounce around between different systems or content providers. It also allows the bank to be more flexible and create unique services assembled around specific events.

Take, for example, the coming US elections. The Marquee team built a new dashboard around the elections using third-party data blended with Goldman analytics and internally-created baskets—say, if taxes go up, what are the implications for certain corporations? Or, will the US have a different relationship with China based on who wins?—as well as content coming from the investment bank’s trading desks. Users can see forecasts for the presidency, as well as House and Senate projections, and they can then build out their own hypotheses. 

And just two weeks ago, the bank launched a similar dashboard specific around environmental, social, and governance (ESG) issues. This dashboard combines analytics from the trading and research departments, along with internal baskets and third-party data to allow users to see in what quartile or decile a company sits from an ESG perspective.

We are also very focused on looking at things that clients would want to outsource to GS that are costly for them to maintain and not providing alpha to their investment process

Anne Marie Darling

Darling is a partner at the firm, managing multi-asset platform sales, global fixed income execution services, and Marquee product and distribution in the bank’s Global Markets division. She says that later this year, or early next, users will be able to filter data based on ESG metrics for the bank’s portfolio optimization tool for equities, as well as for its new credit portfolio tool.

Previously, users would get this information in a half-dozen or more emails from different analysts at Goldman providing their views on the election. The user would have to remember who sent the email, what it said, and then link it on their own to other analytics applications. Now, through these dashboards, that content is right there alongside sales and trading applications to allow the user to more easily create their own themes around a subject and execute on that strategy. 

“Most of our content, historically, in Global Markets was distributed via email,” she tells WatersTechnology. “We need to move that into a much more digital format, where you can search, sign up for alerts, tag certain items, and make it much more user-friendly, similar to what we see in the consumer market. That will be a key driver as we head into 2021.”

Brad Bailey, a research director within research and consultancy Celent’s capital markets division, says that Goldman’s move to open the Marquee platform represents a wider trend seen throughout the industry.

“The more that you can build a platform out and add an ecosystem, where not only you but also your customers and other parties can build into it, that’s tremendous power,” he says. “It’s a fundamentally difficult problem to take all these pieces that have evolved over decades and put them together in a way that’s seamless. That’s the end-game, and it’s kind of a winner-takes-all scenario.”

While Goldman and other large investment banks and asset managers have the challenge of evolving legacy platforms to incorporate cloud-based services and APIs, fintech startups don’t offer the robust, institutional power of a Marquee or similar bank-run platforms.

“Goldman is looking at how their business is changing and how they think about their capital markets, how they think about principle fulfilment, and commissions, and that type of stuff, and they see that in terms of full-service offerings. I think this is everywhere,” Bailey says. “Building natively in the cloud is a lot easier than tying together these systems, but these systems are really, really hard to make. That’s why people want them.”

New Experiences 

Another area where Goldman is building out its workflow capabilities to improve its analytics offering is in foreign exchange. In the first half of 2020, the bank rolled out streaming pricing for swaps. Later this year, it will do the same for options. Once the team completes the options piece, they’ll be able to migrate all users from the legacy Marquee Trader 1.0 platform to the new 2.0 platform, Darling says.

As mentioned previously, Goldman has also been working on a new credit portfolio tool. Instead of a single Cusip—a code that is used to identify securities registered to be sold publicly—clients can access the bank’s roughly 50,000 CUSIPs, similar to what happens in the equity market. This would allow for a more streamlined and efficient workflow for traders, where they can use the same tool that Goldman’s salespeople are using and they can optimize their portfolio for the factors that matter to them—for example, if they want to invest in a specific sector or geography, or remove healthcare companies, or invest only in companies rated AAA, or increase exposure to BBB-rated companies. 

Historically, a client would call an engineer or salesperson, and they’d go back and forth with 20 or so different versions of a spreadsheet to get to the right portfolio. Here, the client or the salesperson can share and iterate on the tool and filter certain Cusips to see what that does to the overall portfolio composition. 

The credit portfolio tool is in beta with a handful of clients, with plans for a broader rollout later this quarter.

“In execution, a lot of our FX users have migrated already to Marquee Trader 2.0; the primary functionality gap to v1.0 in FX options will be closed in the near term, facilitating the migration of all remaining users,” Darling says. “We are very excited about the launch of our new credit portfolio tool, which will allow clients to work with sales and trading, and construct the right portfolio to meet their needs.” 

The New World of APIs, Cloud & Open Source

Marquee has 65,000 monthly active users, Darling says. As the bank has evolved the platform, it has been able to make these changes thanks to the embrace of APIs. She says that everything the Marquee team builds is “API first”, and thus caters to developers to create innovative new tools. 

As a result, any service that can be accessed via the Marquee GUI can also be accessed programmatically. For example, if a developer needs to incorporate data from Goldman for their own web application, the application can make a request for the data via an API. Global Markets has roughly 90 market datasets that it distributes directly to clients from Marquee, Darling says. In this way, the same datasets that are being used by Goldman’s traders to mark their book—say, to mark their volatility surface in equities—can also be accessed by users to back-test an idea. 

Goldman can also offer internal risk tools to users in this way. Darling says that clients often want to access the bank’s pricing and valuation tools, or just Greeks. To allow this, it created GS Quant, a Python tool for quantitative finance on GitHub. Users can tap into the API to develop trading strategies and risk management solutions by manipulating the data provided by the bank. 

The power of combining the content or the idea with engineering and coding is something we have not necessarily done before

Darling

GS Quant allows clients to structure and figure out how to look at pricing holistically, and then look at the Greeks, and sensitivity analysis against the trade structure. And while GS Quant started with Python, the bank is extending it to other languages.

“The power of combining the content or the idea with engineering and coding is something we have not necessarily done before,” Darling says. “And so that is what is really new and exciting as it relates to risk analytics.”  

Coupled with this shift toward APIs, Goldman Sachs increasingly adopting and contributing to the open-source community as the bank looks to create a better experience for developers, which represents a significant shift from how the bank was portrayed during the now-infamous Sergey Aleynikov fiasco. To push this move toward open-source forward, the bank created Developer.GS.com to better reach this audience. Additionally, by tapping into the GitHub community, it can allow users to build upon Marquee themselves. 

“A significant client persona that we need to target is the developers. The developers are going to become much more important for our clients,” she says. 

Tying together these philosophical shifts toward APIs and open source is cloud. Last year, Goldman Sachs announced that it was migrating Marquee onto the Amazon Web Services (AWS) cloud, with the aim of Goldman Sachs taking on processes for clients that do not deliver alpha and are costly, such as ingesting data, sourcing data, or maintaining a security master, though the bank is at the beginning of this exercise, Darling says.

This partnership could represent the seedlings of an eventual managed services offering.

“Moving to the cloud is a multi-year investment for us. We are also very focused on looking at things that clients would want to outsource to GS that are costly for them to maintain and not providing alpha to their investment process. This is what we mean by the creation of the ‘financial cloud’ for our clients,” Darling says.

Examples of financial cloud applications Goldman could develop could include data management and analytics, pricing and risk services, and trading automation. 

Darling continues: “If somebody at one of our clients wants to validate an idea, it is currently very expensive to create and maintain the infrastructure required for quantitative analytics. This includes sourcing and managing high-quality data across different asset classes, building out the infrastructure and models for financial analytics, and maintaining this on an ongoing basis at scale with high accuracy. Goldman Sachs has been doing this for many years, and we are looking to extend these capabilities to our clients. We think that this could be one of our key areas and first steps into the development of the financial cloud.”

As we build out content, we want to consider how can we use machine learning to create a recommendation system for Marquee similar to what Netflix or Amazon does. If you like these five things, you should also be thinking about this.

 Darling

With this data and analytics foundation in place, and now that users can see multiple data sources and trading tools together in one place on Marquee, the next step will be to further incorporate machine learning and other analytics tools so that Marquee can be more proactive in delivering information to users.

“As we build out content, we want to consider how we can use machine learning to create a recommendation system for Marquee, similar to what Netflix or Amazon does,” Darling says. “If you like these five things, you should also be thinking about this. And so we are moving to be proactive with clients about delivering and personalizing things that they value. I think that will be the next evolution for us around content as we head into 2021.”