Tracking inconsistent calculation results reported by users across environments

[Ethann Collins]

Hello everyone,

I'm looking for advice on how others manage and investigate user-reported issues that are difficult to reproduce consistently.

We're working with a web-based calculation platform:

https://scrapgoldmeltcalc.com

The application performs real-time calculations based on user-entered values and has recently generated several support requests reporting inconsistent results.

The challenge is that the issue cannot be reproduced reliably.

Current observations:

• Most users receive expected results.
• A small percentage report different outputs when entering similar values.
• Reports come from different browsers, devices, and locations.
• Some users claim refreshing the page changes the displayed result.
• Support tickets often lack enough technical detail for reproduction.

What we've already done:

• Created Jira issues for each report.
• Added browser and device tracking fields.
• Logged user sessions where possible.
• Reviewed calculation logic and validation rules.
• Compared reported results against expected outputs.
• Checked caching and session-handling behavior.

My questions are:

1. How do you structure Jira workflows for intermittent issues that cannot be reproduced consistently?
2. Do you create separate tickets for each report or consolidate them under a parent investigation task?
3. What information do you require from users before escalating a calculation-related bug?
4. Have you found effective ways to identify patterns when troubleshooting browser-dependent issues?

Since the platform is calculation-driven and relies heavily on user input, we're trying to improve our debugging and triage process before committing development resources to a root-cause investigation.

I'd appreciate hearing how other teams manage similar situations.

Thank you.