DEVOURED
On Software Quality
Software quality depends on user perception shaped by repeated issues and UI/UX experience rather than isolated bugs, requiring symptom-based monitoring of user journeys.
What: An essay arguing that software quality is fundamentally about user perception and trust, which erodes through repeated issues (not single bugs unless catastrophic), and proposing a monitoring strategy focused on "golden paths" through symptom-based metrics backed by system-level alerts.
Why it matters: Challenges conventional defect-counting approaches to quality by explaining why partial outages, transient bugs, and good UX can maintain trust while technically stable systems with poor UX lose it—suggesting teams may be measuring the wrong things.
Takeaway: Map your product features by importance, identify golden paths (core user flows like login/play/exit), monitor them from the client perspective with symptom-based alerts (e.g., "mobile login is slow"), and back those with caused-based system metrics to debug when issues arise.
Deep dive
- Trust in software quality is slow to build but quick to destroy, though modern users tolerate short-lived bugs and only change perception after prolonged or repeated exposure to issues
- Software quality fundamentally means "does this work the way I expect it to," which is subjective—a bug-free app can feel low-quality due to poor UX, while a buggy app with good UX may be forgiven
- Netflix and Facebook strategically had partial outages affecting only some users, so when users asked friends "is it working for you?" the answer was often yes, shifting blame from the platform to the user's setup
- Github, Twitter (fail whale), and others had frequent transient issues but weren't seen as low-quality because problems affected edge cases, were infrequent, or were handled with delightful error UX
- Recent examples show erosion: Github's major outages, Claude Code's repeated availability issues, and Railway's caching incident (leaking user data between accounts) built on weeks of daily incidents
- The ideal metric would be "defects per user per day" weighted by criticality, but this is nearly impossible since you can't categorize bugs you don't know exist
- Symptom-based monitoring focuses on user experience signals (e.g., "login on mobile is slow") rather than component metrics (database query performance), which matters more for quality perception
- Caused-based monitoring tracks the underlying system components; when a symptom alert fires but no caused alert does, it reveals a monitoring gap
- To implement: map all product features, score by importance, identify golden paths (features required to complete core user flows), then monitor each from the client perspective for availability and performance
- Golden paths might be login/play/exit in a game, but not necessarily finding Facebook friends or viewing activity feeds—focus on what's essential to core functionality
- UI/UX is the most important factor in software quality because it's what users interact with; a loading spinner vs blank screen during the same delay creates vastly different quality perceptions
- Good error UX (like Twitter's fail whale or Github's unicorn) can actually make failures more tolerable and even delightful, papering over technical issues
- The dual-layer approach ensures you capture both what users experience (symptoms) and why it's happening (causes), enabling both perception management and root cause fixes
Decoder
- Golden path: The essential user journey through a product—the minimum set of features users need to complete core functionality (e.g., login, perform main action, exit)
- Symptom-based monitoring: Alerts based on user-facing problems like "checkout is failing" rather than underlying technical metrics, capturing quality as users experience it
- Caused-based monitoring: Alerts on underlying system components (database latency, service errors) that cause user-facing symptoms, used for diagnosis and debugging
- CSAT/NPS: Customer Satisfaction and Net Promoter Score—survey metrics that indirectly measure software quality through user sentiment rather than technical defects
Original article
Software quality is driven by user perception—shaped more by repeated issues and UI/UX experience than by isolated bugs—making trust slow to build but easy to erode. To manage this, teams should focus on monitoring user “golden paths” with symptom-based metrics tied to underlying system signals, ensuring they capture both user experience and root causes effectively.
Explore all 66 articles in today's digest
Fresh dev news, AI-digested daily. Never miss a trend at devoured.fyi