Every job I've ever run has had two schedules living side by side. There's the one in the trailer — the big Gantt chart the owner signed off on, taped to the wall, mostly ignored by anyone holding a tool. And there's the real one, the one the drywall foreman and I hash out over coffee about what's actually happening Thursday. The Last Planner System is really just a disciplined way of getting those two schedules to talk to each other, so the wall chart stops being fiction and the coffee conversation stops being improvised.
Software gets pitched as the thing that makes this work. It isn't. The system works because of the conversations and commitments; the software's job is to hold the connective tissue so nothing falls through the cracks between planning levels. Understand the levels first, then you'll know what to demand from any tool you use.
The five levels, and what each one is actually for
Last Planner splits planning into layers that get more detailed and more reliable as the work gets closer. Each layer answers a different question:
- Master schedule — the milestones and the deadline. When does the building have to be dried in, when's the TCO date. This is the promise to the owner. It changes rarely and it should.
- Phase (pull) planning — how a chunk of the job actually gets built, planned backward from a milestone by the people doing it.
- Lookahead planning — the make-ready window, usually three to six weeks out, where you find and kill the constraints before they hit the field.
- Weekly work plan — the commitments for the next week, made by the foremen who'll do the work. Promises, not assignments.
- Daily planning — the morning huddle where the weekly promises turn into "you three are on grid line C this morning."
The mistake nearly everyone makes is treating these as five copies of the same schedule at different zoom levels. They're not. The master schedule predicts. The lookahead prepares. The weekly plan commits. Confuse "we scheduled it" with "we committed to it" and you'll wonder why work you "planned" keeps not happening.
From master schedule down to the lookahead
The master schedule is a lousy tool for running next week and a great tool for orientation. Its milestones need to show up in your lookahead as they approach — not so you can schedule off them directly, but so everyone remembers what the near-term work is feeding. When the crew can see that the pour they're prepping is what lets the milestone hit, the urgency stops being abstract.
What has to flow between these two levels:
- Milestones down. As a milestone comes within a few weeks, it appears in the lookahead as context. It anchors the make-ready work.
- Detail down. A master-schedule bar that says "Level 3 rough-in, 15 days" has to explode into the twenty real activities a lookahead needs — top-out plumbing, in-wall blocking, wall board one side, inspections, insulation. The master bar hides all of that.
- Progress up. When the field closes out those twenty activities, that rolls up to move the master milestone. Do this by hand across two tools and it drifts within a week.
- Variance visible. When the field is tracking behind what the master assumed, you want to see that gap early, while there's still room to react — not at the milestone when it's a crisis.
Here's where software earns its keep or doesn't. If your lookahead lives in a spreadsheet and your master schedule lives in a separate scheduling tool, that progress rollup is a manual re-key every week, and it will be wrong. A platform that keeps the parent-child link between a master activity and its lookahead detail — the way LookAheadWall does when you break an activity into location-based rows — means completion in the field updates the milestone without anyone transcribing anything.
Phase planning: the bridge nobody wants to skip but everybody does
Phase or pull planning is the step teams cut when they're behind, and it's exactly the step that keeps you from getting behind. You gather the trades for a phase — say, the podium deck or a full floor of interiors — and you plan it backward from the milestone. Start at "floor complete," and work in reverse: what has to be done right before that, and before that, each trade putting up their own sticky note or row and stating what they need from the trade ahead of them.
What comes out of a good pull-planning session that you can't get any other way:
- The real handoffs. Framing hands walls to the electrician, who hands them back for inspection, who hands to insulation, who hands to board. Those handoff points are where jobs die, and pull planning makes every one of them explicit and owned.
- The buffers people actually need. When the drywall foreman says out loud he needs a day after MEP rough-in signs off for cleanup and layout before he'll hang, that buffer goes into the plan instead of getting discovered as a fight in week three.
- Long-lead constraints, early. If the pull plan shows switchgear landing in eight weeks and the lead time is sixteen, you just found a problem while it's still solvable.
The output of that session — the sequence, the handoffs, the durations the trades themselves committed to — has to carry forward into your lookahead. Sticky notes on butcher paper are great in the room and gone by Friday. Capturing that sequence as real trade-flow links, so the dependencies survive into the weeks that follow, is the whole point. If pull planning produces a photo of a wall that nobody looks at again, you did the ceremony and skipped the value.
The lookahead: where constraints go to die
This is the level I care most about, because it's where a superintendent actually prevents problems instead of reacting to them. The lookahead — three, four, six weeks depending on your job's lead times — is a screening window. Its one job is to make sure that when an activity reaches the commitment week, it's genuinely ready. Nothing missing.
The discipline is constraint identification. For every activity entering the window, you run the same checklist:
- Materials — on site or with a confirmed delivery date that beats the start?
- Information — RFIs answered, submittals approved, latest drawings in hand?
- Labor — is the crew actually available, or committed to another job that week?
- Equipment — lift, pump, crane time reserved?
- Predecessor work — is the prior trade genuinely done and inspected, not "basically done"?
Every constraint gets an owner and a need-by date. Not "we should chase that" — a name and a date. Then you work them down as the activity moves through the window: constraints get spotted at the far edge, actively resolved in the middle, and verified clear at the near edge. An activity that still has an open constraint does not get to enter the weekly commitment. That gate is the entire mechanism. Break it — let unready work slip into the week because the schedule "says" it's this week — and you're right back to the daily firefighting the system is supposed to end.
A rule of thumb that's saved me repeatedly: size your lookahead to your longest routine lead item, not to a round number. If your typical procurement runs four weeks, a three-week lookahead is structurally too short to make anything ready, and you'll be committing to work you can't supply. This is where a real tool beats a spreadsheet — it flags the activity whose constraint clears after its start date, instead of letting you paper over it.
The weekly work plan: promises, not orders
The commitment week is different in kind from everything above it, and the difference matters. Up in the lookahead, you're screening work. In the weekly plan, the foremen who'll actually do the work look at what's ready and commit to what they'll finish. It's a promise from the person closest to the work, not a task assigned down from the trailer.
That distinction changes behavior. A foreman who publicly commits to closing out three rooms by Friday owns that number in a way he never owns a bar someone else drew. Software that treats these as ordinary scheduled tasks misses the point entirely; the commitment is a separate object with a name on it. The weekly plan is just the sum of those promises, made visible to every trade so the electrician can see what the framer committed to and plan his week against it. That shared visibility is where the coordination actually happens — not in a meeting, in the plan everyone's reading off the same version of.
Daily execution and the huddle
Daily planning is where the week's promises become this morning's work. A ten-minute huddle at the gang box: here's today's push, here's what we finished yesterday, here's what's in our way. You reference the lookahead for what's coming but you're focused on today. The moment someone says "I can't start the east side, the inspector didn't clear the west," that's a threatened commitment — and it needs to surface right then, not Friday when you're tallying up why the week missed. Getting that visibility to a crew leader's phone in the field, rather than trapped on a trailer wall, is exactly what a mobile companion earns its place doing.
Close the loop or the whole thing is theater
Here's the part teams skip, and skipping it is why most Last Planner rollouts fizzle. At the end of each week you measure Percent Plan Complete: commitments completed divided by commitments made. Not "percent of work done" — percent of promises kept. A crew can be busy all week and blow PPC because they did different work than they committed to. That gap is the signal.
Then for every missed commitment, you capture why in plain language. Material late. Inspection slipped. Prior trade wasn't done. Weather. Over a few weeks the reasons cluster, and the cluster tells you where your planning is actually broken. If material constraints cause a third of your misses, your procurement lead times are wrong or your lookahead is too short — and now you know it from data instead of gut. That's the difference between a team that improves on purpose and one that hopes next month goes better. PPC in the 50s early is normal; a healthy team climbs toward the 80s as the make-ready process tightens. The number itself matters less than the trend and the reasons behind it.
What actually connects the levels
The magic of Last Planner isn't any single level — it's that they stay linked. You should be able to trace a task the crew is doing this morning up through the weekly commitment, up through the lookahead activity, back to the phase plan, back to the master milestone. Progress at the bottom flows up. Constraints found anywhere ripple to everywhere. A sixteen-week switchgear lead shows up on the master schedule and tracks all the way down to whether the gear room is a ready commitment in week six.
That traceability is the honest test of scheduling software. Plenty of tools do near-term lookahead planning fine. Fewer keep the thread intact from the master schedule down to a foreman's daily assignment without a human re-keying data between disconnected views. When the levels connect, the wall chart in the trailer and the coffee conversation with the foreman are finally the same schedule. That's the whole game — and it's why we built LookAheadWall around location-based work plans and trade-flow links rather than one more Gantt chart to ignore.
Where to start if this is new
Don't try to stand up all five levels at once. Start with the lookahead and the constraint checklist — that alone will surface problems weeks earlier than you're used to. Once that's a habit, add explicit weekly commitments and start measuring PPC, even if you're just writing the number on a whiteboard. Bring in facilitated pull planning as the team matures, because it takes real skill to run a room full of foremen well. Complete the loop last, but do complete it. Make-ready without measurement gets you halfway; measurement is what turns a good week into a repeatable one.