“Can I use this data feed?”

When a portfolio manager or bank analyst asks this question, a frenzied chain of events occurs involving back-and-forth discussions between the compliance team, a bunch of lawyers, and third-party data vendors trying to decipher data usage rights. There are teams within investment firms that exhaust hours sifting through and interpreting legal jargon in market data contracts.

However, Michelle Roberts, vice president of market data strategy and compliance, and Ilya Slavin, global head of data engineering lead at JP Morgan, say the bank has started to simplify these licensing activities.

Over the last two years, the US investment bank has been working on both internal and industry-led projects to remove inefficiencies in managing and understanding market data contracts. Externally, JP Morgan is a member of the Rights Automation for Market Data Community Group, a group of financial firms formed in January 2020 to build a machine-readable language for market data contracts using the Open Digital Rights Language (ODRL), an open-source data model used for coding policy expression, created by the World Wide Web Consortium.

Industry projects can be slow movers, as they require broad participation, membership consensus, and in the case of ODRL, the rollout of multiple iterations. Slavin says JP Morgan’s team felt the issue of managing data licenses also required more urgent attention, given its impact on the bank’s day-to-day business.

“We want to be able to innovate faster,” he says.

That’s why the bank has been developing a suite of tools to manage licenses. One solution is designed to capture the complete lifecycle of a contract, from signing to renewals and amendments, and all the metadata associated with it. The technology then generates a visualization hierarchy, or as Roberts describes it, a “contract tree,” where the user will see a master copy of the agreement and then multiple peripheral agreements attached to it.

A user such as a compliance team member can search for or query the contracts via the metadata, captured in a large data store. They could look up the date or the timeframe in which the relevant agreement was active.

“The thing that often confounds organizations of every size is that you have to find the right agreement that matches a particular use case, being described for a particular group of users, and that is trying to do something with the data. That is something we focused on first,” Slavin says.

Users will be able to also leverage Google-style text search to look for relevant words or language within the metadata or the digital contract. The text is extracted from the documents using optical character recognition. The tools used to capture the contract lifecycle and make it searchable are currently in production.

Another core part of data licensing is entitlement—who in the enterprise is licensed to access which data. Roberts says the bank is in an advanced stage of developing its own language structure for determining usage rights, and is nearing production of an entitlement tool.

“You select the contract you have, and then use the tool we’ve created to select a series of permissions, constraints, and duties, which are associated with that contract. That then allows you to model the various usage rights in a similar structure to ODRL,” Roberts says.

A user can input into the entitlement tool who they are in the business, what datasets they want to use, and how they want to use them. The tool will let them know if they are or are not entitled to use the data, or flag that there are nuances that might require manual interpretation or other steps to be followed before it grants access to the data.

In a later phase of the build, JP Morgan will develop an engine to turn contract usage rights into ODRL. This could allow users to automate more of the processes around interpreting data licenses, help them to query data more systematically, and interact with other ODRL compliant systems. But achieving this is still some time away, Slavin says.

“The ODRL is an open technology, and open technologies are the glue between different implementations. They can speak the same language, and you do not have the Tower of Babel effect. We’re not at the point where our tech is ready to speak with other parties, but we built all of our systems on internal presentations and are heading toward being fully compliant and eventually completely adapting the ODRL once it is fully ready,” he says.

Marking territory

JP Morgan applied for patents for these data licensing tools two years ago to protect its intellectual property in a crowded marketplace, filing about seven patents with patent protection offices in various countries. The patent application process usually takes about three to four years from filing to approval.

JP Morgan has received a priority date for these patents, which is how the patent office determines which patents were filed first and therefore have priority over others. The patent office will then determine whether the inventions in the application are novel and unique, before JP Morgan’s application is approved.

“We have established the priority dates on a number of different inventions, with our intellectual property claims, and now we have to wait for the machinery of patenting to get to the final state,” Slavin says.

Despite these patent applications, he says, the bank won’t prevent other firms from building similar licensing tools. 

“Patenting is an important part of the process for the development of any new technology. However, we are building these technologies openly and we encourage others to do so as well. That will help the ecosystem,” he says.

JP Morgan has no current plans to sell the tools as commercial offerings, Slavin says, but adds that he can’t speak for the long-term plans of the bank’s product management team.