The Last Planner System gets talked about like it's one thing you turn on. It isn't. It's six moving parts, and I've watched teams bolt on two or three of them, wonder why their Percent Plan Complete is still stuck at 55%, and quietly decide the whole approach doesn't work for their jobs. It works. They just built half a bridge and walked out onto it.
Below are the six components, what each one actually does on a live jobsite, and the specific ways I've seen each one fail. If you're running a look-ahead process now, read this as a diagnostic. Find the component that's weak on your job, because that's the one dragging your reliability down.
Component 1: Master Scheduling — the fence you plan inside
The master schedule (the CPM baseline the GC and owner agreed to) isn't where Last Planner lives, but it's the fence everything else plans inside of. Its job in this system is narrow: give you milestones, phase boundaries, hard external constraints, and the high-level sequence logic. That's it. It is not supposed to tell your electrician what to do on Tuesday.
The thing to get right here is which dates are real. On most jobs, maybe a fifth of the milestones on the master are genuinely hard — permanent power date, owner move-in, a utility shutdown the city scheduled six months ago that will not move for anybody. The rest are float-padded contractor logic. When your field team can't tell the difference, they either treat every date as a fire drill or they stop believing any of them. Walk the master with your supers before the job starts and mark the immovable dates in a different color. Those are what phase planning pulls to.
Where this breaks: the master gets built in the office, printed once, taped to a trailer wall, and never touched again while the field runs on a completely separate reality. Six weeks in, the two schedules have nothing to do with each other. The fix isn't a fancier master schedule — it's a short-interval process that actually feeds progress back up so the master stays honest.
Component 2: Phase Planning — pull planning that earns the sticky notes
Phase planning, or pull planning, is where the trades sit in a room and build the sequence backward from a milestone. Done right, it's the single highest-value hour you'll spend on a project. Done as theater, it's a wall of sticky notes nobody looks at again.
The mechanics that matter: you start at the milestone and pull backward, one trade at a time answering "what do you need in place before you can start, and how long do you need?" Durations come from the foreman who owns the work, not from the scheduler's spreadsheet. And you flush out the handoffs — the exact condition one trade hands to the next. "Framing complete" is not a handoff. "Walls framed, blocking in, top plates fireblocked, and the deck swept" is a handoff, because now the drywaller and the MEP rough-in trades know precisely what they're inheriting.
Two rules of thumb from doing a lot of these. First, put buffer where trades collide, not on every activity. A stack of individually padded durations just hides the real risk and blows your milestone. Concentrate the float at the interfaces — frame-to-rough-in usually wants a day or two for cleanup and inspection sign-off before the next trade opens a wall. Second, the people making the commitments have to be in the room. If a sub sends a project engineer who can't commit the crew, you're negotiating with a messenger and the plan is fiction.
Where this breaks: pull planning happens once at mobilization for the whole job and never again. Do it phase by phase — structure, then envelope, then interior rough-in, then finishes — a few weeks ahead of each phase, when the trades actually know their manpower.
Component 3: Look-Ahead Planning — where you make work ready
This is the engine room, and it's the component most teams do worst. The look-ahead — typically a rolling three-to-six-week window — exists for exactly one purpose: to make work ready before it hits the week it's supposed to happen. You screen every activity entering the window for constraints, you name an owner for each constraint, and you drive those constraints to closed before the activity reaches the front of the line.
The constraints are always the same short list, and you should be hunting all of them: materials and long-lead items on site, approved and returned submittals, shop drawings, RFIs answered, prerequisite work complete, permits and inspections lined up, equipment and access available, engineering and layout done, and manpower committed. If you can't check every one of those boxes for an activity, it is not ready, and it does not belong in a weekly commitment. Full stop.
What good looks like: a six-week horizon where things get more definite as they get closer. Week five is rough — you're just flagging that a switchgear submittal needs to be back. By week two, that constraint is either closed or it's a screaming red item with a name and a date on it. The activities that clear all their constraints drop into what's called the workable backlog — a bench of ready work your weekly plan draws from. That backlog is your shock absorber. When Tuesday's activity gets rained out, you don't send a crew home; you pull ready work off the bench.
Where this breaks: the look-ahead becomes a lie detector nobody runs. Teams "identify constraints" as a meeting formality but never assign an owner or a date, so nothing actually gets made ready, and work still shows up on the weekly plan not ready. Then the crew mobilizes, discovers the material isn't there or the inspection wasn't called, and stands around. Every one of those is a look-ahead failure that happened two weeks earlier. A look-ahead tool like LookAheadWall helps here because the constraint stays attached to the activity and visibly rides toward the current week — but the discipline of naming an owner and a due date is on you, not the software.
Component 4: Weekly Work Planning — commitments, not assignments
This is the heartbeat. Once a week, the people who actually do the work commit to what they'll finish in the coming week — and the word commit is the whole game. A commitment is a promise a foreman makes because he believes he can keep it, made in front of his peers. An assignment is something handed down that he'll nod at and privately ignore. The difference is the entire reason the system produces reliable plans instead of wish lists.
Rules that keep weekly planning honest:
- Only ready work gets committed. If it still has an open constraint, it goes back to the look-ahead. Committing not-ready work is how you train your crews that the plan is meaningless.
- Every commitment is measurable. "Work on Level 3" is not a commitment. "Rough-in plumbing, units 301 through 308" is — you can look at it Friday and say yes or no, with no argument.
- The foreman makes the call, not the scheduler. He knows his crew size Monday, who's out, and what the last job taught him about that detail.
- Handoffs line up across trades. If the drywaller commits to hang Level 2 but the framer's inspection isn't committed until Thursday, you just found a collision in the meeting instead of on the deck.
Where this breaks: the super walks in with a filled-out plan and asks the room to rubber-stamp it. Now they're back to assignments, PPC stays flat, and nobody owns the misses. Let the trades fill in their own lines. Your job is to protect the meeting, keep it to 30 minutes, and make sure the right people are in the chairs.
Component 5: Daily Coordination — keeping the week from drifting
A weekly plan is a promise made Monday about a world that changes by Wednesday. Daily coordination is how you keep the week from quietly falling apart between planning meetings. In practice it's a short standup — ten, fifteen minutes, everyone on their feet — where each foreman says what he finished yesterday, what he's on today, and what's in his way.
The value isn't the status update, it's the "what's in your way." That's your early-warning system. A foreman mentioning he's a hair short on a material or waiting on a layout is a problem you can still solve at 6:45 a.m. The same problem discovered at 2 p.m. with a crew already idle is money out the door. The huddle exists to surface tomorrow's problem while it's still cheap.
Where this breaks: the huddle turns into a status recital with no follow-through, or worse, into a 40-minute sit-down gripe session. Keep it standing, keep it fast, and make sure every "I'm blocked" leaves the huddle with a name attached to unblocking it. This is also where a mobile schedule app earns its keep — the crew leader marks work done and flags a blocker from the deck, and it's visible to the super without a phone-tag relay.
Component 6: Learning — the component everyone skips
Here's the one that separates teams who actually improve from teams who just run meetings. Every Friday you measure Percent Plan Complete: commitments completed divided by commitments made, as a straight percentage. Not "mostly done." Done or not done. A 90%-complete task counts as zero, because the trade waiting on it got nothing.
PPC by itself is just a scoreboard. The gold is in the misses. For every commitment that didn't get done, you categorize why — and there's a standard set of reasons worth using: prerequisite work not complete, materials, information/RFI, labor, equipment, weather, owner/design change, over-commitment, or a change in priorities. Then, over a few weeks, you look for the pattern. If a third of your misses trace to "prerequisite work not complete," your problem isn't the field — it's your look-ahead screening in Component 3 letting not-ready work through. If it's "over-commitment" week after week, your foremen are promising past their crews and need to trim. The variance categories are a map back to the broken component.
Two honest numbers: reliable teams live around 80–85% PPC. A brand-new implementation often opens in the 50s, and that's fine — the first PPC number's only job is to be the baseline you beat. And beware a suspiciously perfect week. A run of 100% PPC usually doesn't mean the crews are superhuman; it means they're sandbagging their commitments to protect the score. That's a coaching conversation, not a celebration.
Where this breaks: teams track PPC, nod at the number, and never touch the variances. Measuring without analyzing is just paperwork. The categorization is the part that pays.
How the six actually connect
These aren't six independent programs — they're a chain, and the whole point is that information flows down and reality flows back up. The master sets milestones. Pull planning turns milestones into a trade sequence. The look-ahead takes that sequence and makes each activity ready. Weekly planning turns ready work into commitments. Daily coordination protects those commitments. And learning measures what actually happened and feeds the lesson back up the chain — usually into whichever component let the miss through.
That feedback loop is why you can't cherry-pick. Skip the look-ahead and your weekly plan fills with not-ready work, PPC craters, and no amount of huddling saves it. Skip the learning and you repeat the same misses every week with great discipline. The components fail as a chain, which is exactly why they only pay off as a chain.
The order to build them in
You don't stand all six up on day one. In my experience the sequence that sticks is:
- Start with weekly commitments. Get the trades making real, measurable promises to each other. This alone changes the culture on a job faster than anything else.
- Add PPC and variance tracking immediately. The week you start committing, start scoring, or you'll never know if it's working.
- Build the look-ahead. Once the team feels the pain of committing not-ready work, formal constraint screening sells itself.
- Layer in pull planning for the next phase, now that they trust the process.
- Reconnect to the master schedule so the field's reality flows back into the baseline.
- Tighten daily coordination last — it's the polish that keeps the week from drifting.
Don't overthink the tooling at the start. You can run this on a whiteboard and sticky notes long enough to prove it works, and plenty of great teams did for years. What software like LookAheadWall buys you once the discipline is real is that the constraint stays stapled to its activity as it rolls toward the current week, the weekly plan is visual and location-based instead of a spreadsheet nobody reads, and the crew leader in the field sees the same plan the super does. The tool doesn't create the discipline. It just stops the discipline from leaking out through re-typed spreadsheets and phone tag.
Where to look first
If your PPC is stuck and you're not sure why, don't reach for a new tool — run the diagnostic. Pull four weeks of variance reasons. Whatever category dominates points straight at the weak component: materials and prerequisites point at the look-ahead, over-commitment points at weekly planning, repeated misses in the same spot point at learning that isn't happening. The six components are only worth anything together, but they break one at a time, and the misses will tell you which one if you bother to read them.