Chapter 27 — Quiz

(c). CPM answers "does the work fit in the time, and what is the longest path?" It assumes readiness — that when a predecessor finishes, the successor can begin. The constraint log answers the readiness question the network cannot see. They are complements, not competitors.

(b). SHOULD (master schedule and pull plan) → CAN (look-ahead and make-ready screen) → WILL (weekly work plan commitments) → DID (PPC and reason codes). (a) is the improvement cycle used inside a first-run study; (c) is 5S with a letter missing.

(d). 0.9⁵ = 0.590. Failures compound multiplicatively across handoffs. Most people estimate somewhere in the 80s, which is exactly why buffers appear everywhere on a job site and nobody can explain where the calendar went.

(b). (a) and (c) are progress and earned-value measures — they belong in the cost report and the EVM report. PPC measures the reliability of the plan, not the quantity of the production.

(d). The screen is absolute. "Probably Thursday" is not removal. Admitting one unready activity idles a crew, displaces it into somebody else's area (manufacturing trade stacking), and — worst of all — teaches the foremen that the plan is decorative.

(c). The eight are: design information, submittals and approvals, material on site, labor available, equipment available, prerequisite work complete, permits and inspections, space and access. Productivity is not a constraint — it is an outcome, and treating it as a constraint is how teams end up blaming crews for a readiness problem.

(c). That is the entire point of the session. Others may ask "what would have to be true for it to be shorter?" — but nobody negotiates a duration down. A duration imposed on a foreman is a number the foreman will not defend, and undefended durations are the reason schedules are ignored.

(b). The last wagon still has to travel through every zone after the first wagon has finished, so the train takes (zones + wagons − 1) takts, not (zones) takts. Nine zones and five wagons is 13 takts, not 9.

(c) — round down. The takt is a drumbeat and every wagon must fit inside it. Rounding up consumes the entire window and leaves no room for the first zone that goes wrong, and a zone will go wrong. Rounding down produces a buffer at the end of the train.

(b). PPC is a ratio and its denominator is controlled by the people being measured. A team that goes from 26 commitments a week at 62% (16 completed) to 9 a week at 89% (8 completed) has halved its real output while improving the headline by 27 points. That is Rivermont Elementary #12, and the job was 17 calendar days behind.

(b). LCI is a nonprofit; "Last Planner" is a registered trademark of LCI. It has no regulatory authority. The research side of the field organizes largely through the International Group for Lean Construction, whose conference proceedings are freely available.

(c). Coordination of prerequisite work, back-scheduling of material need-by dates, and the RFI log are all Kestrel's. That is the normal result, and it is invisible without measurement because general-contractor unreliability shows up as somebody else's crew standing around.

False. They answer different questions. CPM proves the work fits in the time, finds the longest path, supports delay analysis, and communicates with owners. The Last Planner System tells a foreman what is ready on Tuesday. On Northgate both hung on the same trailer wall for six months.

False. 100% means the team is committing only to what is certain — planning below its capacity and leaving production on the table. Sustained figures in the high 70s to high 80s, with a rising commitment count, indicate a healthy system.

False. Three structural differences are permanent: the product is fixed and the workstations move through it; every project is a prototype; and the workforce is assembled from separate companies for a single job. The third is the root of most lean adoption failures, because the party that pays to remove a constraint is usually not the party that benefits.

False. It is a dated, contemporaneous record of who committed to what and what happened. It can be powerful evidence in a delay claim — in either direction. This is also why subcontractors are rationally cautious about writing commitments down, and why a general contractor who uses a PPC log as a claim exhibit destroys that job's data permanently.

False, and this one can cost you a claim. Notice requirements are contractual and jurisdictional, and a planning artifact is not notice. Run both systems. A constraint log tells your team what to fix; a notice letter preserves your rights.

False. It measures the reliability of the plan. Work in place is measured in the cost report and in earned value. Conflating the two is how teams end up reporting excellent PPC on a job that is behind.

Push releases work into the system on a date calculated centrally, whether or not the receiving trade is ready — "the schedule says drywall starts Monday, so drywall starts Monday." Pull works backward from what the downstream trade needs and releases work only when the handoff can actually be received — "in-wall inspection has to pass by Thursday, so framing has to be complete Tuesday, so the frames have to be here Monday." The practical difference is that push produces crews standing in unready areas, and pull produces a shorter list of work that is genuinely available.

The five days were hidden float — his private buffer against a slab edge that might not be ready and against being leapfrogged by the precast crew, which had historically cost him a day and a half per remobilization. He traded it for a guaranteed continuous elevation with nobody else on it.

The lesson: every trade carries hidden float, and they will trade it for reliability — a condition you guarantee — but never for pressure. A general contractor who tries to take the five days by demanding them gets a foreman who says 26 and privately means 30.

1. Protected make-ready time. Test: can you name the person and the hours per week? A job title with no hours means the condition is false.
2. The GC makes and keeps its own commitments, publicly, on the same board. Test: is the GC's own PPC published alongside the trades'?
3. A miss is safe to declare. Test: has anyone been punished — in a meeting, a negotiation, or a claim — for a recorded miss? If yes, the data is already worthless.
4. The commercial terms let the benefit be shared, or the GC knowingly accepts that it is investing. Test: who paid, who gained, and does everyone know?

The near-miss investigation found three failures: a stale competent-person inspection tag, a scaffold modified by a trade that did not erect it, and — the finding nobody wanted to write down — a crew running behind after the steel acceleration under an unwritten pressure to make it up.

The mechanism: variability produces firefighting, and firefighting produces the conditions in which people get hurt. Unready work forces improvisation — night modifications, working in congested areas after being displaced, skipping a tag rather than standing still. A system that surfaces constraints six weeks out generates less of that pressure.

The honest limit: a constraint log would not have re-secured the plank or re-inspected the scaffold. It addresses finding #3, which is the systemic one, and does nothing directly about findings #1 and #2. Claiming otherwise is the kind of overselling that makes practitioners distrust lean.

Diagnosis: adoption as ritual, plus a gamed metric. A stale constraint log means no screening is happening; unscreened work in a plan produces exactly the pattern of high reported reliability and a slipping job — because the commitments have shrunk to things that were already safe.

Two numbers first: 1. Commitments made per week, trended over six months. If it has fallen while PPC rose, the metric is measuring nothing. This takes twenty minutes with the old sheets. 2. Constraints closed per week. If the answer is zero, or nobody can produce it, you have a meeting, not a system.

First three actions: 1. Assign make-ready hours with a name and a source. Twelve to twenty hours a week from a specific person, taken from a specific other duty that you reassign or drop. If you cannot name what you are taking it from, you have not done this step. 2. Add a general-contractor row to the board and score yourself first. Publish your own PPC before you ask anyone for theirs. This is the only fast way to buy back credibility. 3. Re-cut the look-ahead by area and screen it honestly this week. Expect the plan to get shorter. Say out loud that a shorter true plan is the goal, and pre-build a workable backlog so nobody gets sent home.

Then say plainly, in the first meeting, that no recorded miss will be used against anyone — and spend the next four months proving it.