Product development runs on queues you can't see.
Don Reinertsen's book argues that the manufacturing-derived way we manage engineering — keep everyone busy, hit the plan, move work in big phases — is wrong for work whose product is information. In its place he puts 175 principles from economics, queueing theory, statistics, telecom networks and maneuver warfare. Every page here turns them into models you can play with.
The book in one diagram: levers on the left drive queues; queues drive cycle time and feedback; those drive profit. Cost of delay is the exchange rate that tells you how hard to pull each lever.
Hover a node or an arrow. Everything on the left is a lever you control; everything on the right is a result.
Twelve problems with the current orthodoxy
Chapter 1 diagnoses what's broken. Each problem is a belief that works in a factory and fails in a design process — and each one is fixed by one of the chapters that follow.
Track proxy variables — efficiency, schedule, conformance — because nobody can say what a week of delay costs.
Design-in-process inventory is invisible — it's information, not parts — so nobody measures it.
Keep every engineer 100% busy. Idle time looks like waste.
Variability is bad; eliminate it with Six Sigma-style process control.
Success means hitting the original plan, spec and budget.
Phase-gate processes move the whole project as one batch through requirements, design, test.
Meetings, reviews and releases happen when someone asks for them.
Monitor whether activities are on schedule — a lagging indicator.
Start work as soon as it's approved; more started work looks like more progress.
Specialists, fully loaded, each doing exactly their own role.
First in, first out — or whoever shouts loudest gets served first.
Decisions escalate to managers who see the big picture.
Work through the chapters
Turn every trade-off into one currency. Price a week of delay and the rest of the book becomes arithmetic.
Push utilization toward 100% and watch the queue — and your cycle time — go vertical.
Asymmetric payoffs make variability valuable. Shape the payoff, pool the noise, fail cheaply.
Halve the batch and queues, cycle time, risk and feedback delay all shrink with it.
Run the board with and without a WIP limit and compare cycle times, live.
Cadence, synchronization, and WSJF sequencing — the traffic engineering of a dev process.
Delay in a feedback loop makes it oscillate. Shorten the loop and you learn more per year.
Decide where the information is. Intent and decision rules beat escalation queues.
All 175 principles, searchable and filterable, each with a plain-language explanation.
Score your team against the twelve problems and get a prioritized playbook for Monday.
This is an independent study aid, not affiliated with the author or publisher. Principle names follow the book; explanations are paraphrased. Read the book — Celeritas Publishing, 2009 — for the full argument.