C:C02
Chapter 2: Measuring Productivity from Code to Production
A guide to the published second edition, with literal citations separated from wider editorial reading suggestions.
Chapter guide
Examines productivity measures from generated output to accepted releases. It distinguishes randomized evidence from observational comparisons and develops a pilot that records review, rework, quality and hidden human effort before an adoption decision.
Sintesi in italiano
Esamina la produttività dal codice generato ai rilasci accettati. Distingue le evidenze randomizzate dai confronti osservazionali e sviluppa un progetto pilota che misura revisione, rilavorazione, qualità e lavoro umano poco visibile prima di decidere sull'adozione.
Learning goals
- choose a productivity denominator that includes review and rework
- distinguish measured acceptance from reported speed and generated output
- design a comparison that protects against selection bias and missing quality outcomes
Obiettivi in italiano
Obiettivi di apprendimento
- scegliere un denominatore di produttività che comprenda review e rework
- distinguere l'accettazione effettivamente misurata dalla velocità dichiarata e dall'output generato
- progettare un confronto che protegga dalle distorsioni di selezione e dagli esiti di qualità non osservati
Chapter outline
- 2.1 The productivity denominator problem
- 2.2 How to read causal evidence
- 2.3 Why controlled studies disagree
- 2.4 Observational adoption and its confounders
- 2.5 Writing code versus shipping code
- 2.6 Quality-adjusted productivity
- 2.7 Developer experience and hidden time
- 2.8 Designing a coding-specific field pilot
- 2.9 From metrics to an evidence-based pilot decision
- 2.10 Worked engineering example: pick the denominator before the tool
- 2.11 Practice exercise: measure accepted throughput, not generated output
Struttura del capitolo
- 2.1 Il problema del denominatore nella produttività
- 2.2 Come leggere le evidenze causali
- 2.3 Perché gli studi controllati possono essere in disaccordo
- 2.4 L'adozione osservazionale e i fattori confondenti
- 2.5 Scrivere codice non significa rilasciarlo
- 2.6 Produttività corretta per la qualità
- 2.7 Developer experience e tempo nascosto
- 2.8 Progettare un pilot sul campo per task di coding
- 2.9 Dalle metriche a una decisione sul pilot basata sulle evidenze
- 2.10 Esempio ingegneristico svolto: scegliere il denominatore prima dello strumento
- 2.11 Esercizio pratico: misurare il throughput accettato, non l'output generato
Cited in the published chapter
12 linked atlas references from literal citations in the English and Italian editions. General bibliography membership and identifier-index entries are counted separately.
SE@PoliMi bibliography 9
- W001 — AI IDEs or Autonomous Agents? Measuring the Impact of Coding Agents on Software Development
Agarwal · 2026 - W003 — The Effects of Generative AI on High-Skilled Work: Evidence from Three Field Experiments with Software Developers
Cui · 2026 - W004 — Speed at the Cost of Quality: How Cursor AI Increases Short-Term Velocity and Long-Term Complexity in Open-Source Projects
He · 2026 - W008 — How Much Does AI Impact Development Speed? An Enterprise-Based Randomized Controlled Trial
Paradis · 2025 - W009 — Productivity Assessment of Neural Code Completion
Ziegler · 2022 - W010 — Writing Code vs. Shipping Code: Productivity Effects Across Generations of AI Coding Tools
Demirer · 2026 - W020 — Measuring the Self-Reported Impact of Early-2026 AI on Technical Worker Productivity
Becker · 2026 - W023 — Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity
Becker · 2025 - W054 — Not All Agents Are Equal: Code Quality and Post-Merge Maintenance Across Five Autonomous Coding Agents in the Wild
Kraishan · 2026
Addy Osmani 2
- O029 — The reality of AI-Assisted software engineering productivity
Addy Osmani · 2025-08-16 - O049 — The 70% problem: Hard truths about AI-assisted coding
Addy Osmani · 2024-12-04
The Pragmatic Engineer 1
- P061 — How tech companies measure the impact of AI on software development
Gergely Orosz; Laura Tacho · 2025-09-16
Trace literal citations to chapter sections
Locations come from the supplied English and Italian source files. Chapter text and figure captions are identified separately; identifier-index references and general bibliography membership are not treated as chapter citations.
- O029 · The reality of AI-Assisted software engineering productivity
- EN · Writing code versus shipping code
EN/chapters/chapter02.tex:100 - ITA · Scrivere codice non significa rilasciarlo
ITA/chapters/chapter02.tex:100
- EN · Writing code versus shipping code
- O049 · The 70% problem: Hard truths about AI-assisted coding
- EN · Writing code versus shipping code
EN/chapters/chapter02.tex:100 - ITA · Scrivere codice non significa rilasciarlo
ITA/chapters/chapter02.tex:100
- EN · Writing code versus shipping code
- P061 · How tech companies measure the impact of AI on software development
- EN · Designing a coding-specific field pilot
EN/chapters/chapter02.tex:155 - ITA · Progettare un pilot sul campo per task di coding
ITA/chapters/chapter02.tex:155
- EN · Designing a coding-specific field pilot
- W001 · AI IDEs or Autonomous Agents? Measuring the Impact of Coding Agents on Software Development
- EN · Observational adoption and its confounders
EN/chapters/chapter02.tex:78 - ITA · L'adozione osservazionale e i fattori confondenti
ITA/chapters/chapter02.tex:78
- EN · Observational adoption and its confounders
- W003 · The Effects of Generative AI on High-Skilled Work: Evidence from Three Field Experiments with Software Developers
- EN · How to read causal evidence
EN/chapters/chapter02.tex:34 - ITA · Come leggere le evidenze causali
ITA/chapters/chapter02.tex:34
- EN · How to read causal evidence
- W004 · Speed at the Cost of Quality: How Cursor AI Increases Short-Term Velocity and Long-Term Complexity in Open-Source Projects
- EN · Observational adoption and its confounders
EN/chapters/chapter02.tex:78 - EN · Quality-adjusted productivity
EN/chapters/chapter02.tex:113 - ITA · L'adozione osservazionale e i fattori confondenti
ITA/chapters/chapter02.tex:78 - ITA · Produttività corretta per la qualità
ITA/chapters/chapter02.tex:113
- EN · Observational adoption and its confounders
- W008 · How Much Does AI Impact Development Speed? An Enterprise-Based Randomized Controlled Trial
- EN · How to read causal evidence
EN/chapters/chapter02.tex:39 - EN · Why controlled studies disagree
EN/chapters/chapter02.tex:49 - ITA · Come leggere le evidenze causali
ITA/chapters/chapter02.tex:39 - ITA · Perché gli studi controllati possono essere in disaccordo
ITA/chapters/chapter02.tex:49
- EN · How to read causal evidence
- W009 · Productivity Assessment of Neural Code Completion
- EN · The productivity denominator problem
EN/chapters/chapter02.tex:19 - ITA · Il problema del denominatore nella produttività
ITA/chapters/chapter02.tex:19
- EN · The productivity denominator problem
- W010 · Writing Code vs. Shipping Code: Productivity Effects Across Generations of AI Coding Tools
- EN · The productivity denominator problem
EN/chapters/chapter02.tex:19 - EN · Writing code versus shipping code
EN/chapters/chapter02.tex:92 - ITA · Il problema del denominatore nella produttività
ITA/chapters/chapter02.tex:19 - ITA · Scrivere codice non significa rilasciarlo
ITA/chapters/chapter02.tex:92
- EN · The productivity denominator problem
- W020 · Measuring the Self-Reported Impact of Early-2026 AI on Technical Worker Productivity
- EN · Developer experience and hidden time
EN/chapters/chapter02.tex:134 - ITA · Developer experience e tempo nascosto
ITA/chapters/chapter02.tex:134
- EN · Developer experience and hidden time
- W023 · Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity
- EN · Why controlled studies disagree
EN/chapters/chapter02.tex:49 - EN · Developer experience and hidden time
EN/chapters/chapter02.tex:134 - ITA · Perché gli studi controllati possono essere in disaccordo
ITA/chapters/chapter02.tex:49 - ITA · Developer experience e tempo nascosto
ITA/chapters/chapter02.tex:134
- EN · Why controlled studies disagree
- W054 · Not All Agents Are Equal: Code Quality and Post-Merge Maintenance Across Five Autonomous Coding Agents in the Wild
- EN · Quality-adjusted productivity
EN/chapters/chapter02.tex:113 - ITA · Produttività corretta per la qualità
ITA/chapters/chapter02.tex:113
- EN · Quality-adjusted productivity
Editorial reading mappings
Broader topical suggestions can overlap the cited sources. A mapping does not add a citation to the published edition or establish that a practice is effective.
SE@PoliMi bibliography 50
- W001 — AI IDEs or Autonomous Agents? Measuring the Impact of Coding Agents on Software Development
Agarwal · 2026 - W002 — Code with Me or for Me? How Increasing AI Automation Transforms Developer Workflows
Chen · 2026 - W003 — The Effects of Generative AI on High-Skilled Work: Evidence from Three Field Experiments with Software Developers
Cui · 2026 - W004 — Speed at the Cost of Quality: How Cursor AI Increases Short-Term Velocity and Long-Term Complexity in Open-Source Projects
He · 2026 - W005 — Early Adoption of Agentic Coding Tools by GitHub Projects
Raida · 2026 - W006 — Agentic Much? Adoption of Coding Agents on GitHub
Robbes · 2026 - W007 — Usage, Effects and Requirements for AI Coding Assistants in the Enterprise: An Empirical Study
Merler · 2026 - W008 — How Much Does AI Impact Development Speed? An Enterprise-Based Randomized Controlled Trial
Paradis · 2025 - W009 — Productivity Assessment of Neural Code Completion
Ziegler · 2022 - W010 — Writing Code vs. Shipping Code: Productivity Effects Across Generations of AI Coding Tools
Demirer · 2026 - W011 — The Productivity-Reliability Paradox: Specification-Driven Governance for AI-Augmented Software Development
Farrag · 2026 - W012 — Agentic Coding in the Wild: Characterizing GitHub Copilot Traces at Production Scale
Liu · 2026 - W013 — Quantifying the Expectation-Realisation Gap for Agentic AI Systems
Lobentanzer · 2026 - W014 — Adoption and Impact of Command-Line AI Coding Agents: A Study of Microsoft's Early 2026 Rollout of Claude Code and GitHub Copilot CLI
Murphy-Hill · 2026 - W015 — Agentic Delegation and the Language Frontier of Software Developers: A Model and Evidence from Claude Code on GitHub
Quispe · 2026 - W016 — The Impact of AI Coding Assistants on Software Engineering: A Longitudinal Study
Vella · 2026 - W017 — Randomized Controlled Trials for Conditional Access Optimization Agent
Bono · 2025 - W018 — Intuition to Evidence: Measuring AI's True Impact on Developer Productivity
Kumar · 2025 - W019 — Generative AI and labour productivity: a field experiment on coding
Gambacorta · 2024 - W020 — Measuring the Self-Reported Impact of Early-2026 AI on Technical Worker Productivity
Becker · 2026 - W021 — The AI Engineering Report 2026: The Acceleration Whiplash
Faros AI · 2026 - W022 — AI Coding Agents: Adoption Trends
JetBrains Research · 2026 - W023 — Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity
Becker · 2025 - W024 — 2025 State of AI-Assisted Software Development
Google Cloud DORA · 2025 - W025 — Estimating AI productivity gains from Claude conversations
Tamkin · 2025 - W026 — Accelerate State of DevOps Report 2024
Google Cloud DORA · 2024 - W027 — Beyond the Copilot: Scaling the Agentic Product Development Life Cycle
Relyea · 2026 - W028 — The AI Revolution in Software Development
Relyea · 2026 - W033 — How AI Coding Agents Modify Code: A Large-Scale Study of GitHub Pull Requests
Ogenrwot · 2026 - W044 — An Empirical Study of Code Clones from Commercial AI Code Generators
Wu · 2025 - W051 — From Conversation to Contribution: Characterizing Coding Agent in Open-Source Software
Fang · 2026 - W054 — Not All Agents Are Equal: Code Quality and Post-Merge Maintenance Across Five Autonomous Coding Agents in the Wild
Kraishan · 2026 - W055 — Debt Behind the AI Boom: A Large-Scale Empirical Study of AI-Generated Code in the Wild
Liu · 2026 - W057 — How Do AI Coding Agents Contribute to Software Development? An Empirical Study of Agentic Pull Requests
Mazloomzadeh · 2026 - W059 — AI-Generated Smells: An Analysis of Code and Architecture in LLM- and Agent-Driven Development
Zhu · 2026 - W074 — RovoDev Code Reviewer: A Large-Scale Online Evaluation of LLM-based Code Review Automation at Atlassian
Tantithamthavorn · 2026 - W084 — Self-Preference Bias in LLM-as-a-Judge
Wataoka · 2024 - W095 — Beyond Code Generation: Reliability, Verification, and Cost Economics in the Agentic Software Development Lifecycle
Bhati · 2026 - W108 — Do LLM Evaluators Prefer Themselves for a Reason?
Chen · 2025 - W117 — The Augmentation Trap: AI Productivity and the Cost of Cognitive Offloading
Caosun · 2026 - W118 — How AI Impacts Skill Formation
Shen · 2026 - W124 — Artificial Intelligence, Productivity, and the Workforce: Evidence from Corporate Executives
Baslandze · 2026 - W145 — ROI of AI-Assisted Software Development
DORA · 2026 - W146 — McKinsey Technology Trends Outlook 2026
Chui · 2026 - W163 — A Case Study of Onboarding in Software Teams: Tasks and Strategies
Ju · 2021 - W177 — Stuck and Frustrated or in Flow and Happy: Sensing Developers' Emotions and Progress
Müller · 2015 - W182 — Don't Touch My Code! Examining the Effects of Ownership on Software Quality
Bird · 2011 - W185 — Succession: Measuring Transfer of Code and Developer Productivity
Mockus · 2009 - W187 — Globalization by Chunking: A Quantitative Approach
Mockus · 2001 - W188 — The Ramp-Up Problem in Software Projects: A Case Study of How Software Immigrants Naturalize
Sim · 1998
Addy Osmani 2
- O029 — The reality of AI-Assisted software engineering productivity
Addy Osmani · 2025-08-16 - O049 — The 70% problem: Hard truths about AI-assisted coding
Addy Osmani · 2024-12-04
The Pragmatic Engineer 22
- P008 — IDEs with GenAI features that Software Engineers love
Gergely Orosz · 2024-11-26 - P014 — How AI-assisted coding will change software engineering: hard truths
Gergely Orosz; Addy Osmani · 2025-01-05 - P021 — Cross-platform mobile development with Expo
Gergely Orosz; Elin Nilsson · 2025-02-04 - P031 — What is Growth Engineering?
Gergely Orosz; Alexey Komissarouk · 2025-03-20 - P038 — What’s Changed in 50 Years of Computing: Part 4
Gergely Orosz · 2025-05-06 - P042 — Microsoft is dogfooding AI dev tools’ future
Gergely Orosz · 2025-05-27 - P047 — Learnings from two years of using AI tools for software engineering
Gergely Orosz; Birgitta Böckeler · 2025-06-24 - P048 — Software engineering with LLMs in 2025: reality check
Gergely Orosz · 2025-07-01 - P050 — The Pragmatic Engineer 2025 Survey: What’s in your tech stack? Part 1
Gergely Orosz; Elin Nilsson · 2025-07-15 - P052 — Cursor makes developers less effective?
Gergely Orosz · 2025-07-24 - P056 — The Pragmatic Engineer 2025 Survey: What’s in your tech stack? Part 2
Gergely Orosz; Elin Nilsson · 2025-08-19 - P061 — How tech companies measure the impact of AI on software development
Gergely Orosz; Laura Tacho · 2025-09-16 - P078 — Frictionless: why great developer experience can help teams win in the ‘AI age’
Gergely Orosz (introduction/interview); Nicole Forsgren; Abi Noda (excerpts) · 2025-12-10 - P083 — Inside a five-year-old startup’s rapid AI makeover
Gergely Orosz · 2026-01-27 - P085 — Deepdive: How 10 tech companies choose the next generation of dev tools
Gergely Orosz · 2026-02-03 - P089 — AI Tooling for Software Engineers in 2026
Gergely Orosz; Elin Nilsson · 2026-03-03 - P091 — Are AI agents actually slowing us down?
Gergely Orosz · 2026-03-17 - P100 — Revisiting “No Silver Bullets” in the age of AI
Gergely Orosz · 2026-05-12 - P107 — Slow down to speed up: so much has changed in 6 months’ time
Gergely Orosz · 2026-06-23 - P112 — How building software is changing at Anthropic
Gergely Orosz · 2026-07-28 - P121 — Why has Shopify dropped React Native?
Gergely Orosz · 2026-09-29 - P123 — Cycles of disruption in the tech industry: with software pioneers Kent Beck & Martin Fowler
Gergely Orosz (credited author and moderator); Kent Beck and Martin Fowler (panel participants) · 2026-04-07
Martin Fowler’s website 6
- F003 — In-line assistance - when is it more useful?
Birgitta Böckeler · 2023-08-01 - F020 — Partner with the AI, throw away the code
Matteo Vaccari · 2025-07-31 - F029 — Experiences with local models for coding
Birgitta Böckeler · 2026-07-08 - F030 — The Economic Benefit of Refactoring
Giles Edwards-Alexander · 2026-07-30 - F033 — Are Harnesses Worth the Cost?
Naresh Jain · 2026-10-13 - F035 — When Software Meets Hardware
Kuldeep Singh · 2026-10-15
Earlier Non-Deterministic books 12
- B003 — Summary
Enrico Papalini · - B009 — Measuring What Matters---The Data Behind AI's Real Impact
Enrico Papalini · - B016 — PART IV: Organizational Foundation
Enrico Papalini · - B018 — Antipatterns: What Not to Do
Enrico Papalini · - B020 — Understanding Tool Maturity Levels
Enrico Papalini · - B029 — Measurement \& Observability
Enrico Papalini · - B031 — First 90-Days Implementation Roadmap
Enrico Papalini · - B032 — Quick Reference Guides
Enrico Papalini · - B033 — Expert Insights
Enrico Papalini · - B043 — Limitations and Judgment
Enrico Papalini · - B064 — Limitations and Professional Judgment
Enrico Papalini · - B106 — The Economics of Loop Engineering
Enrico Papalini ·