How Thomson Reuters mastered Master Data Management and delivered business benefit

Profile picture for user Mark Samuels By Mark Samuels May 21, 2021 Audio mode
The media giant’s data chief explains how his team is overcoming the challenges of a complex technology and business environment to deliver insight-led services to customers.

Thomson Reuters logo

Technology leaders who want to make the most of the disparate data that their organizations hold must find a way to bring information together in a structured but flexible manner that meets the clear requirements of business users.

That’s the opinion of Marc Alvarez, Vice President for Data Management and Operations at Thomson Reuters, who has spent the past two-and-a-half years building a strategic Master Data Management (MDM) programme alongside technology partner Winshuttle and its EnterWorks platform that is closely aligned to his organization’s broader goals.

Alvarez explained to attendees at Gartner’s European Data & Analytics Summit how building the MDM platform involved a complex challenge because of a range of business and technological variables. However, he said the benefits of pursuing the initiative are clear:

We now feel we have a strong view of data content. It's all inventoried, we've built a catalogue around it, we've built our automation for data quality management around it, and all of that is not only supporting the business looking backwards but it's really about driving our business forwards and positioning ourselves for the future.

Mastering complexity

Alvarez said the scale of the challenge his team faced was due to a range of factors. First, Thomson Reuters is a big company, with a $35 billion market cap and more than 500,000 customers. Inevitably the company’s IT architecture and its data structures have became more complex as the business has expanded:

I think this is something that's common across the economy and many firms. We arrived at a point where we literally had dozens of billing and CRM and other workflow systems across the company to drive our business. That causes a lot of conflicts.

While pockets of good data governance existed around Thomson Reuters, there was no agreed organization-wide standard. That lack of governance constrained the amount of analytics and business intelligence the firm could generate from data, which is a problem for the company’s business users who want to serve clients as effectively as possible. Alvarez explained: 

They're suffering from very inconsistent views, with sometimes conflicting views of customers and products – it’s very difficult to reconcile pricing to product, country and business unit. And these domains continue to grow as we move forward.

One final challenge was that the company split into two parts in 2018 – Thomson Reuters and financial data specialist Refinitiv, which is now owned by the London Stock Exchange. The split helped elucidate the need for effective data management, said Alvarez:

How we actually manage our whole lifecycle with our customers really required us to take a step forward – a very complicated step – of consolidating all our customer- and product-related information. The means for achieving that is by building better business intelligence capabilities – and doing it at scale across the company.

The aim was to create a common technology platform across the company’s data sets. He said better governance – in the form of ensuring all the data components come together as a single service – became a key element of success. Achieving such governance makes it easier to reach scale, grow quickly and respond effectively, according to Alvarez:

We were very convinced very early that we needed to work with a partner who could help us and equip us to achieve those goals.

Implementing the platform

Thomson Reuters undertook a detailed requirements-definition program. This process highlighted the need for a multi-domain approach that could draw-in data from across the business, that was cloud-based, and on a platform that was highly configurable, so users could take advantage of data without having to rely on additional coding.

Alvarez and his team used these requirements to create a proof of concept that focused on product and pricing data, which was a complicated area given the company has about 700,000 product skews. They build a range of models and iterated them, before creating a rationalization strategy that mapped all data to a single logical model across the platform. This logical model provided a clear view of products and services:

We now have the ability to actually navigate that content. And that's quite useful for things like standardised product-pricing tables across geographies, where the only thing that differs is currency, not necessarily the price of the product. So these are the types of flexibility and agility we were shooting for.

The team then undertook a thorough RFP process that assessed vendors across a range of parameters from integration to performance and onto pricing. In the end,  Winshuttle and its EnterWorks platform were selected and the implementation process got underway, with the first data domains rolled out within six months. Alvarez said: 

We’ve standardized our product and pricing domain. We’ve built out our business contacts domain, which involved integrating data from 30 sources into one master view. And we've rebuilt our customer master platform, so that we've now brought it into a much more agile configuration-based approach.”

There is now a sense of predictability in place. When Thomson Reuters wants to address additional data domains, it has an awareness of scope, timelines and budget. As for  best-practice lessons for other data leaders from the project, Alvarez advised: 

You can't do enough on the requirements side. Even after we finished publishing our first draft of documentation to the company, we found that stimulated a lot more discussion about what else could be done and what other demands should be supported. You have to view your requirements spec as a living document.

Thomson Reuters is now starting to offer a lot of products and services through customer self-service that were previously unavailable. The firm is also generating other insights, such as data on product use and the effectiveness of online marketing. This evolution is important, said Alvarez, arguing that data leaders must avoid thinking of MDM as a one-off project:

“his is all about being pro-active in ensuring that the data is created correctly, authored correctly, published correctly, and then it meets the needs of the business applications. One of the advantages we have is we've taken this modular, domain-based approach and built outwards, so we weren't looking at this as a single silo. And I think you really need to do that.