Every framework in the Last Planner System eventually funnels down to one document: the weekly work plan. The master schedule sets the destination, the phase pull plan sets the route, and the six-week look-ahead clears the road. But none of that moves a single stud into place. The weekly work plan is where the whole system either pays off or falls apart, because it's the one place people stop talking about what should happen and start promising what will.
I've watched crews run beautiful pull-planning sessions with color-coded sticky notes covering an entire wall, then walk out to a jobsite that looked nothing like the plan by Wednesday. The gap was almost never the strategy. It was the weekly plan — vague commitments, work that wasn't actually ready, and a Monday meeting that had quietly turned back into a foreman reading assignments off a Gantt chart. So let's talk about how the weekly plan actually works when it's done right, and where it goes sideways.
What the Weekly Work Plan Is Actually For
The weekly plan isn't a to-do list. A to-do list is something a superintendent hands down. A weekly work plan is a set of promises the trade foremen make to each other, out loud, in the same room. That distinction is the entire ballgame, and I'll come back to it.
Functionally, a good weekly plan does four things at once. It records who is going to complete what, and by which day. It surfaces the handoffs — the moments where one trade finishes and the next one can start — so nobody is standing around Thursday morning wondering why the deck isn't ready. It confirms the crew, the material, and the equipment are actually lined up behind each promise. And it gives you a clean measuring stick at the end of the week, so you can tell whether your planning is getting better or you're just busy.
Strip away the jargon and it comes down to this: reliable weekly plans are how a project stops surprising itself.
Only Ready Work Belongs on the Plan
Here's the rule that separates real short-interval scheduling from wishful thinking: if the work isn't made ready, it doesn't go on the weekly plan. Full stop.
Made-ready means the constraints are gone. The design question is answered, the RFI came back, the material is on site or has a confirmed delivery, the preceding trade is genuinely done (not "basically done"), the permit is posted, the crew exists. The look-ahead's whole job over the preceding weeks is to knock those constraints down one at a time so that by the time an activity is a candidate for the weekly plan, it's clean.
When you commit to work that still has an open constraint, you're not planning — you're gambling, and you're teaching your crews that the plan is fiction. I've seen foremen learn within a month that half the weekly plan won't survive contact with the jobsite, at which point they stop taking it seriously and start freelancing. Once that happens, you've lost the system. Protect the plan by keeping unready work off it, and keep a small backlog of ready, low-dependency filler tasks so a crew that gets blocked has somewhere productive to go instead of standing down.
Commitments, Not Assignments
The traditional model pushes work down: here's what the schedule says you owe this week. The Last Planner model pulls it the other way: given what's actually ready and what you actually have, what can you promise to finish?
This sounds like a semantic nicety until you watch it change behavior. When a foreman is assigned a task and it slips, the failure belongs to the schedule — "the schedule was unrealistic." When a foreman commits to a task in front of his peers and it slips, the failure is a broken promise, and everyone in that room saw him make it. That social weight is the engine. It's why the commitment has to come from the person doing the work, not from the super narrating a printout.
The practical tell that you've slipped back into assignments: the superintendent is doing most of the talking in the planning meeting. If you're reading the plan to the foremen instead of the foremen reading it to you, you've got assignments wearing a Last Planner costume.
What a Real Commitment Looks Like
A commitment you can actually track and score has to be specific enough that anyone walking the floor could tell whether it's done. "Work on electrical" is not a commitment. "Rough-in receptacles and switches, units 101 through 108, complete and ready for inspection by Thursday" is.
Before it goes on the plan, run each commitment through a quick filter:
- Scoped: A definite quantity in a definite location, not a vague direction.
- Verifiable: "Done" means the same thing to the person committing and the person inspecting.
- Dated: Tied to a specific day, not just "sometime this week." Vague timing hides slips until it's too late to react.
- Resourced: The crew size, the material, and the equipment behind it are confirmed — not assumed.
- Sound: Every constraint is cleared. This is the made-ready check, applied one more time at the moment of commitment.
Foremen who are new to this will over-promise for the first few weeks — it feels good to sound aggressive in the meeting. Let the numbers correct them. Nothing teaches realistic commitment like watching your own miss rate on Friday.
The Coordination That Happens in the Room
The single most valuable thing that comes out of a weekly planning session isn't the plan itself — it's the ten minutes where the framer and the mechanical foreman discover they both planned to own the same corridor on Wednesday. Weekly planning is where trade-flow conflicts get caught while they're still cheap to fix.
So work the seams deliberately. For every commitment, ask what has to be finished before it can start, and who owns that predecessor. Ask whether two crews are landing in the same space at the same time — the classic drywall-versus-overhead-MEP collision, or flooring going in while punch is still touching the walls. Ask about shared resources: the one man-lift, the single hoist, the loading dock window. And ask whether sequence matters within the week, because "both by Friday" is not a plan if one has to precede the other by two days.
This is exactly where a visual, location-based look-ahead tool earns its keep. When you can see the whole week laid out by area with each trade's flow drawn across it, the overlaps jump off the screen instead of hiding inside a list. LookAheadWall was built around that view specifically so the space and sequence conflicts are obvious before anyone commits, not after someone's already mobilized.
Running the Weekly Session
Keep the meeting tight and keep it on a rhythm. Ninety minutes, same time every week, no exceptions — the moment sessions start getting skipped or rushed, reliability craters. A structure that works:
- Look back (15–20 min): Score last week. Calculate PPC, and for every miss, capture why. This goes first on purpose — it sets an honest tone before anyone makes new promises.
- Update the look-ahead (20–30 min): Walk the constraint log. What got cleared, what's still blocked, what's newly ready. Confirm the pool of made-ready work.
- Plan the week (30–40 min): Foremen commit, in their own words. Coordinate the handoffs. Confirm resources against each promise.
- Close out (5–10 min): Read the commitments back, confirm they're captured, assign any follow-ups, and get the plan out the door before people leave the room.
Attendance is non-negotiable. You need the superintendent leading, the trade foremen who can actually speak for their crews, a project engineer to catch action items and chase information, and sub representatives who have the authority to commit for their company. A foreman who has to "check with the office" before promising anything isn't a decision-maker, and the meeting stalls waiting on people who aren't there.
Measuring It: PPC and Why It's Not a Grade
Percent Plan Complete is the simplest metric in the whole system and the most misused. It's just:
PPC = commitments completed ÷ commitments made × 100
A commitment is either 100 percent done or it's a zero. There's no partial credit — "80 percent framed" is a miss, because the successor trade can't start on 80 percent of a wall. That harshness is deliberate; it keeps people honest about what "done" means.
As a rough read: consistently above 80 percent means your make-ready process is working and your commitments are realistic. Seventy to eighty is workable but leaking — you've got constraint gaps sneaking through. Below 70 and something is systematically broken, usually either unready work getting onto the plan or foremen over-committing under pressure.
The trap is treating PPC as a report card and beating people up over the number. Do that once and your PPC will magically climb — because foremen will start committing to almost nothing, or padding "done" to protect the score. The number becomes a lie and you've destroyed the one honest signal you had. PPC is a thermometer, not a hammer. A crew that commits aggressively and hits 75 is worth more to you than one that sandbags to a comfortable 95.
The Part That Actually Improves the Project: Variance Analysis
The PPC number tells you that you missed. The reasons tell you why, and the why is where all the improvement lives. Every incomplete commitment should get a cause logged before people leave Friday, while memories are fresh. Sort them into consistent buckets — prerequisite work not done, materials, information/RFI, labor, equipment, weather, over-commitment, change in priority.
One miss is noise. The pattern across four or five weeks is the gold. If a third of your misses trace back to "predecessor not complete," your problem isn't the weekly plan at all — it's upstream make-ready or a specific trade that keeps declaring victory early. If materials dominate, your procurement lead times don't match your look-ahead window. If "information" keeps showing up, your RFI turnaround is too slow to feed the plan, and you need to be logging those constraints weeks earlier.
This is the plan-do-check-act loop that makes the system compound. You commit, you build, you score, you find the pattern, you fix the process, and next week's plan is a little more reliable than last week's. Do that for a couple of months and the improvement is real and measurable — the whole point of running this instead of a static bar chart.
Getting the Plan Into People's Hands
A plan that lives on the superintendent's laptop coordinates nobody. The instant the session ends, the plan needs to be out — pushed to phones, posted in the trailer, whatever it takes. Each foreman should be able to see their own commitments at a glance and see the overall project view showing how their week interlocks with everyone else's.
The field piece matters more than people admit. When crews and crew leaders can pull the current plan up on a phone and mark progress from where they're standing, the plan stays alive all week instead of dying the moment it's printed. That's the role the mobile companion plays — a crew leader sees exactly what their crew committed to and updates it in real time, so Friday's scoring reflects the actual jobsite instead of a guess reconstructed from memory.
The Failure Modes to Watch For
Almost every weekly-plan breakdown I've seen traces back to one of a short list of sins:
- Dictated commitments. Assignments dressed up as promises. If the super is doing the talking, this is happening.
- Vague scope. Commitments too fuzzy to score, which conveniently means nobody ever technically fails.
- Unready work on the plan. The original sin. Constraints not cleared, promises built on hope.
- Missing coordination. Individually fine commitments that collide in space or sequence because nobody worked the handoffs.
- Skipped or rushed sessions. The discipline decays the moment the meeting becomes optional.
- Weaponized PPC. Using the number to punish, which teaches everyone to game it.
Notice that most of these are prevented upstream — by honest make-ready discipline in the look-ahead — not fixed at the weekly meeting. By the time bad work reaches your Monday session, it's already too late to make it good.
Where the Software Fits
None of this requires a tool. Plenty of great teams run the Last Planner System with sticky notes, a whiteboard, and a disciplined superintendent, and if that's working for you, don't let anyone talk you out of it. What software buys you is leverage: it captures commitments with the required detail so nothing vague slips through, it flags space and sequence conflicts before they're mobilized, it tallies PPC and sorts variance reasons automatically so the pattern is visible without a spreadsheet, and it gets the plan onto every phone the second the meeting ends.
That's the honest case for a purpose-built short-interval scheduling app. It doesn't make the commitments for you and it won't clear your constraints — that's still your crew's work. But it removes the friction that causes teams to quietly abandon the discipline after a few busy weeks, which is how most Last Planner rollouts actually die.
Master the weekly work plan and the rest of the system has something to converge on. Get the commitments honest, keep unready work off the board, work the handoffs in the room, score it straight, and learn from every miss. Do that consistently and the project stops surprising you — which, after twenty years of jobsites, I've come to think is the only kind of predictability that's real.