Sources

Research and Industry Evidence Ledger

Clickable list with the exact origin of each quantitative point used across the site.

This website combines peer-reviewed studies, preprints, project pages, and industry materials. Peer-reviewed findings are treated as highest confidence. Project and industry sources are labeled and used as practical context.

SourceAuthorsVenue / YearKey finding usedTypeLink
What Makes a Good Bug Report?Nicolas Bettenburg, Sascha Just, Adrian Schröter et al.FSE / technical report, 2008Developers report a recurring information gap: missing steps to reproduce, stack traces, test cases, environment data.peer-reviewed paperhttps://www.st.cs.uni-saarland.de/publications/details/bettenburg-tr-2008/
Information Needs in Bug Reports: Improving Cooperation Between Developers and UsersSilvia Breu, Rahul Premraj, Jonathan Sillito, Thomas ZimmermannCSCW, 2010Sampled 600 bug reports and documented early-stage clarification questions in bug handling workflows.peer-reviewed paperhttps://thomas-zimmermann.com/publications/files/breu-cscw-2010.pdf
Deep Learning Based Valid Bug Reports Determination and Automated Bug TriageJ. He et al.ISSRE, 2020Large-scale study and model for classifying valid/invalid reports using text-only features.peer-reviewed paperhttps://doi.org/10.1109/ISSRE5003.2020.00026
Who Should Fix This Bug?John Anvik, Lyndon Hiew, Gail C. MurphyICSE, 2006Triaging baseline of about 5 minutes per report was used to benchmark triage workload.peer-reviewed paperhttps://dl.acm.org/doi/10.1145/1134285.1134336
Revisiting the Performance of Automated Approaches for the Retrieval of Duplicate Issue ReportsRakha et al.EMSE, 2017Reported duplicate prevalence in literature ranges that motivate 20–30% baseline duplicate burden.peer-reviewed paperhttps://researchr.org/publication/RakhaBH18
An Empirical Study of Bug Fixing RateWeiqin Zou, Xin Xia et al.COMPSAC, 2015Mozilla and Eclipse fixed/non-fixed outcomes used as non-fixed quality indicators.peer-reviewed paperhttps://research.monash.edu/en/publications/an-empirical-study-of-bug-fixing-rate/
Feedback-Driven Automated Whole Bug Report Reproduction for Android AppsX. Zhang et al.ISSTA, 202490.63% success, 74.98 sec/report in benchmark.peer-reviewed paperhttps://arxiv.org/abs/2407.05165
Prompting Is All You Need: Automated Android Bug Replay with Large Language ModelsAuthors listed in arXiv submissionarXiv, 202481.3% success with 253.6 sec/report.peer-reviewed paperhttps://arxiv.org/abs/2306.01987
BugCraft: End-to-End Crash Bug Reproduction Using LLM Agents in MinecraftEray Yapıcı et al.ASE, 2025Manual crash reproduction baseline of 3.41 days was reduced to minutes in several agent-comparison cases.peer-reviewed paperhttps://arxiv.org/abs/2503.20036
Enhancing Mobile App Bug Reporting via Real-Time Understanding of Reproduction StepsMattia Fazzini, Kevin Moran, Carlos Bernal Cardenas et al.IEEE Transactions on Software Engineering, 2022Report authors claim reporters built reports ~31% faster with guided S2R capture and higher reproducibility quality.preprinthttps://arxiv.org/abs/2203.12093
Automated Generation of High-Quality Bug Reports for Android ApplicationsBugScribe authorsarXiv preprint, 202644.1%–82.3% relative improvement in S2R completeness and 3.8%–35.2% improvement in observed/expected behavior quality.preprinthttps://arxiv.org/abs/2604.01148
Automated Duplicate Bug Reports DetectionKang LiDiVA thesis/industrial experiment, 2017Industrial experiment context reported estimated value around 2.8 hours per duplicate and 39,019 total projected hours in the referenced dataset context.preprinthttps://bth.diva-portal.org/smash/record.jsf?pid=diva2%3A1153748
BugScribe project pageBugScribe TeamProject page, 2026Project materials report a manual writing drop from roughly 20 minutes to under 2 minutes.project page

Scenario assumption context included

https://bug-scribe.github.io/
Adopting Automated Bug Assignment in Practice: A Longitudinal Case Study at EricssonMarkus Borg, Leif Jonsson, Emelie Engström et al.IEEE/Empirical Software Engineering, 2024Reported industrial adoption context with automation-driven assignment, including speedups and assignment accuracy context.industry reporthttps://arxiv.org/abs/2209.08955
The QA-Developer Handoff: Why 35% of Sprint Time Is WastedSnagRelay article authorsindustry article, 202635% sprint-time waste claim and follow-up rate drop from >60% to under 10%.industry reporthttps://snagrelay.com/
What Is Support Cost per Ticket in SaaS?Industry benchmarking sourceindustry benchmark, 2025Average support ticket cost used as operational planning benchmark, $15.56 in provided findings.industry report

Scenario assumption context included

https://www.which-50.com/support-cost-per-ticket-statistics/
Bug Report Optimization: Creating Reports That Get Fixed FasterIndustry authorsindustry article, 2025Reported benchmarked closure from 13 days to 4 days with higher report quality.industry reporthttps://fullscale.io/blog/bug-report-optimization-distributed-teams/
The Hidden Tax of Incomplete Bug ReportsQAFlow case study authorsindustry case study, 202618–22 hours/week clarifications; 2.7 back-and-forth rounds; 8–12 hours delay per round.industry case studyhttps://www.qaflow.com/