Help Chains

Definition

A help chain is a preplanned protocol defining what counts as a problem, how someone signals they need help, who responds first, how quickly help must arrive, and what happens if the first responder can't resolve it — escalation continues up a defined chain until the issue is closed. In manufacturing this is the andon system: an operator flips a light, a team leader is obligated to respond immediately, and if they can't fix it within the line's cycle time, the problem escalates automatically. Oosterwal argues lean systems need help chains precisely because they run with little slack — without a fast, defined path to assistance, the first unresolved problem stalls the whole system.

In the Book

Chapter 14 recounts how Harley-Davidson's product-development team discovered the gap the hard way. Their first pilot of "pull events" (regular convergence checkpoints) worked well because the small number of projects involved got close oversight — any issue was caught and escalated immediately, which meant the team had unknowingly built a help chain into the experiment through sheer attentiveness. When they tried to scale pull events across the whole project portfolio, that close oversight became impossible, the informal help chain disappeared with it, and the new system nearly failed — not because the underlying cadence-and-flow logic was wrong, but because no formal escalation mechanism existed to keep the system flowing when (not if) problems surfaced. The book contrasts manufacturing, where a stopped line is immediately visible, with product development, where "it is very difficult to see that the process has stopped flowing" — which is exactly why an explicit, defined help chain matters more in knowledge work than on a physical line, not less.

Why It Matters

Any process run near capacity — a support queue, an on-call rotation, a lean workflow with tight WIP limits — needs an explicit answer to "who do I ask, and how fast do they have to respond" before the first blocker occurs, not improvised in the moment. The lesson generalizes past manufacturing: informal escalation that works at small scale (because everyone happens to be watching closely) silently disappears as scale grows, and only a designed, preplanned chain survives that transition.