The Product Development Limit Curve¶
Definition¶
When you plot the timing of a project's design-test-redesign cycles across its calendar, the data forms a consistent curve — steep near launch, flat early on — and a project's position relative to that curve, not its progress against phase-gate checkpoints, predicts whether it will launch smoothly or spiral into crisis. Projects that stay to the curve's "safe zone" (redesign loops completed within the timing the curve allows) launch cleanly; projects pushed to the "risk zone" rarely recover.
In the Book¶
The discovery (Chapters 8–9) came from Oosterwal revisiting years of Harley-Davidson launch data with his mentor Allen Ward after noticing zero correlation between hitting phase-gate exit dates and project success — sometimes the correlation ran backward. Re-plotting the same data by the timing of test-and-fix redesign loops instead produced an R² of 0.99: near-perfect correlation. Allen recognized the shape immediately: "You have a limit curve." The book traces why the curve has its shape — early in a project, few deadlines create little urgency and design alternatives are still wide open, so the first redesign loop takes a long time; near launch, the organization throws every resource at problems because sunk costs are enormous and schedule has become king, so loops close fast. Quality VP Steve Phillips's complaint — "How the hell is it possible to get a month behind in two days?" — captures exactly the blind spot the curve exposes: teams look fine against gate checklists right up until they hit a redesign loop the curve says they don't have time left to close.
Why It Matters¶
The limit curve replaces a fragile leading indicator (did we hit the checklist milestone) with one grounded in the actual mechanism that determines outcomes (how much redesign-loop time is left versus how much is needed). Any project or process with iterative rework — software, hardware, drug trials — can be diagnosed the same way: track the cadence of your fix-verify loops against a launch deadline, not just phase completion, because loop timing is what actually predicts whether you converge or spiral.