Trusted records help Western Union deliver better customer experiences to improve the bottom line

Why knowing 90% of your customers isn’t enough

To better serve customers, Western Union is transitioning from a retail-first organization – in the U.S. alone there are 57,000 Western Union agent locations – to a digital-first one. As part of this transformation, the financial services company is integrating records from digital and retail channels to form 360-degree customer views.  

But when 1.2 billion people have used your services to send and receive money in the last decade and 375 million in the past two years via multiple channels (retail stores, a website, point of sales transactions, Interactive Voice Response, and a mobile app) – data volume, variety and quality issues make obtaining those single customer views challenging. 

“You may have five profiles but you are the same person. And I should be able to give you the same service whether you're using the mobile app, the retail channel or the website. We want to bring all of that information together and offer a great omni-channel engagement,” said Harveer Singh, Chief Data Architect, Western Union. But without accurate customer data, providing an elevated level of service ”was absolutely impossible.”

Western Union did have some insight into their customers. They used a 10-year-old rules-based master data management tool that cleaned around 80% to 90% of the customer data. However, that wasn’t enough to impact the bottom line.

“You make a real difference in that 10% to 20% zone. That’s when you can reduce fraud or increase revenue. Our product has remained the same so to have a business result, we have to look at how much money we can make from those customers or how much risk we can reduce,” he said.

Conventional rules-based systems can be effective on a small scale, relying on human-built rules logic to generate master records. However, rules quickly fall apart when tasked with connecting large amounts of highly variable data (like millions of customer records) at scale.

“To master 375 million records using the traditional way of writing rules, we’d be doing it for 10 years and probably still wouldn’t be able to do it,” Singh said.

Adopting a modern approach to MDM

To adopt a modern approach to MDM, Western Union deployed Tamr Mastering. With Tamr’s next-generation solution, machine learning handles a majority of the manual data preparation work. Humans are kept in the loop to provide feedback and boost confidence in the machine learning model.

"The way mastering is done, our data analysts and data scientists were blown away by it. Most of their time was spent writing rules and here they were building experiences for our customers while bringing Tamr along that journey to build that heart of our data foundation,” said Singh.

Western Union also uses Tamr to enrich records with external data to standardize mailing address formats, validate phone numbers, and check that email addresses have valid domains. Combining data enrichment with data mastering means that not only are data sources automatically cleaned, they are also enhanced with valuable commercial information while avoiding the extremely time-consuming and manual work that goes into incorporating external data.

Each customer record is also assigned a Tamr ID, a persistent identifier that serves as a reference to an entity across multiple data sources assigned by Tamr. The customer records created by Tamr are loaded into the Snowflake Data Cloud, providing Western Union with a single source of truth to power both analytical and operational applications.

“Tamr is becoming the heartbeat of our organization because that Tamr ID gets propagated to downstream systems as well as front end systems,” Singh said. “Having customer records that we know are right helps us get more value from Snowflake.”

Western Union runs Tamr on its AWS instance, allowing them to take advantage of the platform’s flexibility and scalability. This combination of using AWS, Snowflake and Tamr together fits with Western Union’s plan of using modern technologies to better serve customers.

“We don’t want to be in the hardware business. We want to be in the business of building products for customers. With AWS and Snowflake, we have the flexibility and scalability of the cloud and less hardware costs and administration. Tamr offers machine learning so we don’t spend time managing a lot of rules. Using them all together helps us with our ultimate goal of understanding our relationships with our customers,” he said.