When Sift was founded in 2011, we were one of the first companies to apply AI to combat payment fraud. At the time, this was revolutionary—using machine learning, a subset of AI, to detect patterns that humans couldn’t see was a game-changer for digital businesses struggling with fraud. But over a decade later, I’m struck by a realization that might surprise some: single point-in-time transaction data alone is insufficient for truly effective fraud prevention.
The reality is counterintuitive but powerful. Only about 23% of the signals in Sift’s Global Data Network come from transaction points. The vast majority of data that powers our fraud decisions comes from other critical parts of the consumer journey.
Why the Entire Consumer Journey Matters
Think about the last time you shopped online. Before you ever entered your credit card information, you likely browsed products, created an account or logged in, searched for specific items, added them to your cart, and perhaps even abandoned that cart once or twice. Each of these touchpoints creates valuable signals that, when analyzed collectively over time, paint a comprehensive picture of user behavior.
Fraudsters follow a similar routine outside of the checkout page, leaving digital footprints with every move they make. By the time a transaction occurs, those footprints are the key to stopping cyberthiefs.
Building Identity Trust Through Better Signals
Sift champions the concept of “identity trust”—the likelihood that a digital user or interaction can be trusted. Identity trust isn’t established in a single transaction; it’s built from accumulated and consolidated data points across the consumer journey and over a time horizon.
What does this mean in practice?
- When a user first visits your site, their device information, browsing patterns, and navigation behavior all contribute valuable signals.
- Account creation and login activities provide critical identity markers and potential risk indicators.
- Search behaviors and content interactions reveal patterns distinctive to legitimate users versus fraudsters.
- Cart activities provide behavioral context that transactions alone cannot.
- Post-purchase behaviors complete the picture with return patterns and customer service interactions.
By analyzing these behaviors across the 1 trillion annual events within our Global Data Network, we’re able to establish patterns that can help you to distinguish legitimate customers from potential threats with a high degree of accuracy.
The Data-Driven Foundation of Effective Fraud Prevention
For businesses still relying primarily on transaction-point fraud prevention, consider this: you may be missing approximately 77% of the available data signals that could help you make more informed decisions. It’s inefficient at best, and can leave existential vulnerabilities in your fraud operations.
Our data scientists have consistently found that the most predictive fraud signals often emerge well before the transaction. A user who exhibits unusual browsing patterns, creates accounts with suspicious velocity, or interacts with your site in ways that deviate from established patterns will have a greater likelihood of being identified by our systems before they attempt to complete a fraudulent purchase from a Sift customer.
Moving Beyond the Transaction-Only Mindset
Despite the clear advantages of full-journey fraud prevention, many businesses remain trapped in a transaction-focused paradigm. This is understandable, as transactions are where financial risk is most immediately visible. But this narrow focus creates blind spots that sophisticated fraudsters have become experts at exploiting.
The businesses winning the fraud prevention battle today are those that are able to make fraud decisions in-the-moment, with an array of signals that a consumer or fraudster has left over time not just on one site or app, but across an entire network of platforms. They’re establishing identity trust by providing stronger protection from fraud before it ever reaches the assessing risk at multiple touchpoints, providing stronger protection from fraud before it ever reaches the transaction stage.
The Future of Fraud Prevention is Journey-Based
As fraudsters continue to evolve their tactics, the importance of comprehensive journey data will only increase. At Sift, we’re constantly refining our models to identify new patterns across the consumer journey.
For digital businesses facing increasingly sophisticated fraud threats, the message is clear: transaction data alone isn’t enough. The path to more effective fraud prevention—and to building confidence in identity trust—runs through transactional and behavioral signals across journey stages and over time.
