In the new world, it is not the big fish which eats the small fish, it’s the fast fish which eats the slow fish.
Experian, the world’s largest credit bureau and a global information services provider, wants to be one of those rare companies that can be both a big fish and a fast fish, and is evolving its mainframes and monolithic databases into a modern set of cloud technology to make that happen.
For the company, its newly created platform-as-a-service and consequent entry into the API business could mean increasing growth and relevance in a hypercompetitive market. For the millions of consumers and businesses it serves, Experian believes the new technology platform will help it deliver more precise credit decisions, better protection against fraud and identity theft, and faster access to the information necessary to grease the wheels of modern digital commerce.
Platform for a data business, not a credit bureau
Technology has always been a driving force in Experian’s history, shaping it from a small firm advising local merchants to a $4.6 billion company serving clients in more than 80 countries.
Experian’s business today is separated into four areas – credit services for businesses, decision analytics, data marketing services and direct-to-consumer services. It competes with both large players such as Equifax and TransUnion, as well as smaller, more nimble startups – especially in newer areas like data services.
In competitive markets like these, Experian’s growth and differentiation depends on developing new data services and expanding into new markets and regions as soon as an opportunity presents itself. As anyone who has worked at a large, global company can attest, this kind of nimbleness is not easy.
To ensure it can capitalize on these new opportunities, Experian chose to develop its own platform-as-a-service called Experian Data Fabric, on which it could quickly build and deliver products and services that share information. The modern architecture would create a more stable environment for managing petabytes of data, enable the company to run more complex models and analytics, and provide a faster and more accurate customer experience.
Today the platform is made up of a number of cloud and open source components – Cloudera for data, Apache Kafka and DataTorrent for ingestion, and Okta for identity management to name a few. The applications that sit on the platform are written in Java and can be run using Heroku, Amazon Web Services or Experian’s internal infrastructure.
This is the type of stack one would expect from a start-up, not a company that’s been around for 125 years. Merv Lally, Experian’s senior vice president and chief enterprise architect, explains what motivated this move from mainframes to multi-tenant cloud:
Experian is a technology, software, and data company. In the past we built monolithic applications that were cumbersome to change and expensive to maintain. Our vision is to evolve to a build once and deploy anywhere….We call it cloud native development. Even if it’s not deployed to the cloud, it still needs to be transportable.
That vision is making a big impact on the company already.
What took me six months to do, now I can do it in six hours. I’ll give you an example. We’re modernizing our bureau in Columbia. There’s about 40 million consumers from Columbia, and let’s say an average consumer has about seven accounts associated with them. You do the math. We were able to bring in all those records and calculate the [credit] states in six hours on the first try. When we did that with a previous bureau, it took us almost 6 months. That’s a big, big change.
It also improves stability. Lally compares it to the classic block-stacking Jenga game:
Imagine all the Jenga pieces are events that make up your credit. When you first put the Jenga tower together, it’s a very stable structure. That’s your credit. But if there’s an error in one of those events, the system process finds the break, pulls it out of the middle and replaces it with a new event on top. You do that multiple times and you can see how unstable the structure around your credit gets. It’s hard to manage. It becomes hard to do calculations on. It affects downstream processes.
With the architecture and modern components of the new platform, the system can quickly knock down the structure, take out the pieces it doesn’t want, put in the new pieces and build it back up for a more stable structure every time.
Experian joins the API economyNow that the Experian platform-as-a-service has put a common data fabric in place, it is embarking on another significant project. One that gets the company into the data-as-a-service game.
Over the last year, Experian has been working on an API program that will allow business customers to interface with its array of data sources in a customized way, so they can pull exactly the information they need in real-time. These APIs should also speed product development for both Experian and its business customers.
Traditionally Experian would build products that each customer would log into separately, resulting in duplicate data entry, processing delays, integration costs and creating a frustrating user experience. Now Experian’s product team can build one product with an API and provide customers access to those APIs so they can easily embed them within their own products and services. Merv provides an example of how this provides a better customer experience:
With banks, we had to send files back and forth. We processed the files and sent the results back as an ingestion file. Now, all a bank has to do is call the API and send their data through that. Our system processes it, and sends it back in milliseconds. This gives financial institutions real-time decisioning, real-time fraud detection. We get less false positives, you get a happier customer.
Uber is another example. The global ride-sharing company uses Experian’s data to check the credit background of potential drivers and to validate information on riders such as address, email, phone number, and so on.
Experian first started talking about using APIs as a new channel for revenue about a year ago. The technology team partnered closely with the business around program direction and to determine which APIs to prioritize – typically based on customer interest and revenue impact.
The company went live in November after an extended beta period to get people used to this new channel, and the first round of revenue-generating APIs went live this month (February). More than 1,000 developers have registered for Experian’s API program, and the company is working with a handful of its more forward-thinking large customers in the financial services, healthcare, telecom and auto industries. It’s still early days, but this could be a game changer for how IT works with the business.
Putting identity front and center
Security and identity management are critical for any business, but it’s high stakes for a company like Experian that handles personal and financial information for nearly a billion people worldwide. As the company begins to make more APIs available, which expose data in a different way, the stakes become even higher.
Lally and his team knew they needed a single, reliable identity management layer that could unify the architecture and replace the hodgepodge of identity systems it currently used throughout the company.
When I got to Experian, we did not have one standard solution. We were using products from RSA, Ping Identity and Oracle, MobileIron for mobility management, Okta for one of our consumer credit apps, and probably a handful of others. We evaluated all of them for which had the best success and best cost of ownership.
The team ultimately went with Okta. The fact that Okta was one of the first cloud-based solutions for identity management helped the case given Experian has taken a cloud-first approach to its internal systems. Experian has since deployed Okta to all 17,000 employees and is now in the process of converting Experian’s clients. Today it is up to about 1 million external users.
Standardizing on a single solution has brought cost savings as well. Lally estimates he will be able to get rid of six different technologies, eliminating six different service contracts and nearly 30 servers. He expects to recover the full cost of Okta after about 14 months, and a year of that time has already gone by.
The rise of disruptors like Uber and Netflix has taught large companies an important lesson. They can’t rely on their size or existing market dominance to maintain their lead. The ones who move faster, win. The pace of the digital world demands it.
Startups get this. It why they were some of the earliest adopters of the cloud and didn’t bother building everything from scratch. Large global companies can’t typically be that nimble. Many are saddled with decades of technology debt. They stagger under the weight of their own bureaucracy, and can be filled with long-tenured employees who bring a mindset of, “This is the way it’s always been done, it’s too difficult to change now.”
It’s refreshing to see a company that’s been around as long as Experian look past these limitations and use technology to push the company and the culture forward. Merv Lally will be the first to tell you it’s not an easy process, but one that’s necessary if you want to stay relevant.