Fans of Sherlock Holmes know that the great detective solved cases through keen observation, incisive reasoning, and encyclopedic knowledge. AI is nowhere near deerstalker status, but modern technology’s ability to absorb and analyze huge amounts of data while looking for patterns of fraud means that, with the right tools and the right database-driven back end, retailers can be equipped with considerable crime-fighting capabilities.
“We're transforming the way companies fight online fraud and protect their customer accounts,” says Zac Rosenbauer, Vice President of Technology at Precognitive. “We provide multi-dimensional fraud prevention technology that combines device intelligence, advanced behavioral analytics, and a real-time decision engine to accurately detect and prevent fraud for online businesses.” This means capturing and analyzing vast amounts of data from around the world, in near-real-time.
We're transforming the way companies fight online fraud and protect their customer accountsZac Rosenbauer, Vice President of Technology
Real-time Fraud Detection
Precognitive is a Chicago-based online fraud detection firm acquired by ShopRunner, an e-commerce network offering top retailers enterprise technology products including same-day delivery exchange and marketplace services to grow their business while elevating customers’ digital shopping experiences. The range of fraudulent activity Precognitive scans for includes CNP, or card-not-present crime where the fraudster presents someone else’s card details, and ATO (account takeover), where the fraudster logs in as someone else. Depending on the customer, Precognitive also prevents other cybercrime related activities such as bot attacks, payment fraud, and return abuse.
Real-time fraud detection needs to be fast, scalable, flexible and above all reliable. In online retail, downtime is more than just an immediate hit on the bottom line. The competition is a click away, and frustrated customers know it.
Built on MongoDB Atlas and Google Cloud
“We built our system on MongoDB Atlas and Google Cloud,” says Rosenbauer. “I'd be adamantly opposed to moving away from either.” Although Precognitive started with MongoDB’s self-hosted community in 2016, a recent move to Atlas helped the company raise its game. “There’s so much we no longer have to do now that we’re on Atlas.” Rosenbauer says. “We really like the built-in backups, multi-regionality, auto-scaling, and Atlas Charts. We’ve got clients on three continents, but with hosted MongoDB we can serve them easily without the cost of dedicated staff doing database management.”
Although the hard-cost savings of moving to Atlas in raw infrastructure terms have been minimal, not hiring a DBA is a big benefit, Rosenbauer says. And the real win has been the increased productivity of the rest of the eight-person development team, which doesn’t have to worry about major database decisions. He notes that, with the support of the MongoDB Atlas team, “It’s been awesome: no maintenance issues. We upgraded from version 3.4 to version 4.2 in two weeks. In retrospect, we could have done it in two days."
The Technology Roadmap Is at the Heart of Customer Satisfaction
“We sit in a critical path for most of our customers, so we have to deliver. The reliability of MongoDB has allowed us to re-invest in product development and really hone in on supporting our clients. We are investing in the product, not database management.”
MongoDB’s close integration with Google Cloud Platform and its natural affinity for JSON makes it easy to bind into the company’s service stack, says Rosenbauer. “We’re a Node.js-based organization. Our technology stack is completely built on top of Google Cloud, using Kubernetes, Kong as our API gateway, Bigtable as our data store for behavioral and device data, and MongoDB Atlas as our database for everything else.”
Because of the company’s comfort with Atlas’ security management, he says, Precognitive is moving as much as possible onto the platform to minimize the management required. “We are able to meet our DRP (disaster recovery planning) objectives due to the automated back-ups, ability to query those back-ups and the overall reliability of the system. This has allowed us to handle availability incidents with ease.” The company’s roadmap includes many of MongoDB’s newest products. “The new features introduced in Atlas 4.2 are going to be a big focus for us in 2020. For example, Atlas Search allows us to have a built-in full-text search engine for our data. We’re planning to move lots of our search functionality, possibly all of it, to Atlas Search. In addition, materialized views are our number-one priority. Lightning-fast reporting is a huge bonus when you’re managing a global fraud detection environment. All this makes Atlas more than just a database, it’s a capable platform that gives us many new angles to improve the reliability and speed of our fraud analysis.”
“We’re heavily invested in MongoDB Atlas,” Rosenbauer says. “It’s core to our technology stack. We’ve made three-plus years of investment in the database, and we really appreciate not having to move away from it to accomplish our plans. It’s right at the heart of our customer satisfaction.”