Case Study
Data-Driven Error Reduction
The Challenge
A data import process was generating hundreds of errors daily. The team was spending significant time handling these case-by-case, treating each error as a unique problem to solve.
The Approach
Rather than treating symptoms, I analyzed the error patterns using database queries. By joining data from multiple sources and grouping errors by frequency, I identified the true distribution of problems — something that wasn't visible when handling errors one at a time.
The Result: 75% of all errors stemmed from just 1-2 root causes. Two targeted fixes eliminated the majority of the problem — work that would have taken weeks of case-by-case handling.
The fix wasn't more error handling code. It was asking the right questions of the data.
Case Study
Workflow Friction Causing Incomplete Data
The Challenge
A stakeholder was frustrated that users weren't entering complete records. The assumption was that this was a training or compliance issue — users just weren't following the process correctly.
The Discovery
After meeting directly with users, a different picture emerged. The data entry workflow had multiple friction points that made entering complete records tedious and time-consuming. Users weren't cutting corners out of carelessness — they were making rational choices to save time in a system that worked against them.
The Approach
Rather than adding more training or enforcement, I worked with the users themselves to identify the biggest pain points in their workflow. We prioritized fixes based on impact and implemented changes to streamline the process.
The Result: Complete record rates increased significantly, and downstream processes that had been slowed by missing data began running smoothly.
When users aren't following a process, the instinct is often to blame the users. But the real question is: what is the system making difficult?