Bypassing patent databases entirely to find a predictive UI algorithm in a defunct 2003 developer forum.
A massive e-commerce platform was sued by a Non-Practicing Entity (NPE) over a patent claiming a method for predicting cart abandonment based on user mouse movements.
The patent claimed novelty by tracking "cursor hesitation" and trajectory away from the checkout button, triggering a pop-up discount before the user could close the tab.
Cut-off Date: Prior to 2006
Software patents often describe algorithms in broad, functional legal terms rather than specific code. Searching for "cart abandonment" in early 2000s patent databases yielded zero technical matches.
While academic papers discussed predictive UI, they didn't use commercial e-commerce terminology. The exact combination of "mouse tracking" and "triggering a pop-up" was missing from formal literature.
We stripped away "e-commerce" and "shopping cart." The core logic was simply: tracking cursor hesitation and trajectory to predict user intent.
We bypassed patents entirely. We used the Wayback Machine and legacy code repositories (like SourceForge) to search for early 2000s web analytics tools.
We located archived developer discussions where early open-source creators shared raw code snippets to solve UI problems before commercial software existed.
We tracked down a defunct, open-source web analytics startup from 2003. In an archived developer forum, the creator published the exact algorithm under the name "cursor-hesitation tracking."
The code was freely shared to help webmasters see where users got confused and trigger a "help" pop-up. The logic mapped 1-to-1 with the plaintiff's "novel" e-commerce claims.