Every superintendent has stood in the trailer staring at a wall of Gantt bars that told a beautiful story about a job that no longer existed. The bars were baselined six months ago. Since then the steel came late, the owner changed the lobby finish twice, and the mechanical sub pulled two crews to chase a fire somewhere else. The schedule on the wall still says drywall starts Monday. Everybody in the room knows it doesn't. That gap — between the schedule and the jobsite — is the whole argument between traditional CPM scheduling and the Last Planner System.
This isn't a story about one being right and the other being garbage. They answer different questions. Understanding which question you're actually trying to answer this week is how you stop wasting time defending a document nobody believes and start planning work that actually gets built.
What each one is actually for
Traditional Critical Path Method scheduling is a prediction engine. You feed it activities, durations, and logic ties, and it computes when things should happen, what the critical path is, and where you have float. It's contractual, it's strategic, and it's genuinely good at answering big questions: if this milestone slips two weeks, what dies downstream? Do we have the float to absorb a late permit? When do we need the tower crane gone so the next building can start?
The Last Planner System is a commitment engine. It doesn't try to predict July from a desk in January. It works in the window you can actually see — the next few weeks — and it asks a different question: of the work we say we'll do this week, which of it is truly ready, and who is promising to complete it? The output isn't a prediction. It's a set of promises made by the people who have to keep them.
Here's the trap most jobs fall into: they own a beautiful CPM schedule and then try to run the field off it directly. CPM was never built to run a Tuesday. It's too coarse, it updates too slowly, and it assumes constraints resolve themselves. The field needs something with a shorter fuse.
Prediction versus commitment, on the ground
The difference shows up in language. A CPM-driven job says "the schedule shows drywall starting the 14th." A Last Planner job says "the drywall foreman commits to closing zones A and B this week." One is a statement about a document. The other is a promise a human being made in a room full of the trades he depends on.
That distinction matters because predictions fail quietly and commitments fail loudly. When a baseline date slips, nothing happens until someone runs an update three weeks later and discovers you're behind. When a weekly commitment fails, you know Friday morning, in the plan review, and you ask the only question worth asking: why? Was the area not ready? Did the material not show? Did the crew get poached? That "why" is where the improvement lives.
The constraint problem nobody plans for
If I had to name the single biggest difference between the two approaches, it's how they treat constraints. Traditional scheduling assumes them away. The logic tie says electrical rough-in finishes before drywall, so the schedule confidently marches on as if the rough-in will be done, inspected, and signed. On paper it always is. On the job, half the time it isn't — the inspector didn't come, one panel is waiting on a part, the low-voltage guy never showed to drop his conduit.
Last Planner treats those as constraints to be hunted down and cleared before the work is ever put on a weekly plan. This is the part of the system that earns its keep, and it lives in the look-ahead. In a rolling three-to-six-week look-ahead you take every activity coming up and screen it against the same short list of questions:
- Prerequisite work — is the predecessor actually complete and inspected, not just "mostly"?
- Materials — is it on site or firmly scheduled, with a confirmed delivery date, not a hope?
- Information — are the drawings current, the RFI answered, the submittal approved?
- Manpower and equipment — is the crew committed and the lift/scissor/hoist available that day?
- Access and conditions — is the area released, safe, and physically ready to work?
Anything that fails a screen becomes a tracked constraint with a name and a due date attached to it — not "someone should handle the elevator inspection" but "Mike is clearing the elevator inspection by Thursday." Work does not move onto a weekly work plan until it's constraint-free. That single discipline is why crews on a well-run Last Planner job spend their days building instead of standing around discovering the area isn't ready.
A practical rule of thumb: if you can't clear a constraint in the six-week window, it doesn't belong in the six-week window yet — it belongs on a longer-range make-ready list where somebody senior is chasing it. The look-ahead is for work you can realistically make ready, not a wish list.
The metric that actually changes behavior
Traditional scheduling measures variance: how far actual dates drifted from baseline dates. It tells you that you're eleven days late. It does not tell you why, and it does not tell you what to do Monday.
Last Planner measures Percent Plan Complete — of the commitments you made for the week, what percentage did you actually finish. If you planned forty tasks and completed thirty-two, you're at 80% PPC. That number is brutally honest and immediately useful. A crew running at 50% PPC isn't lazy; it's being handed work that isn't ready, and the PPC number forces that conversation. A job climbing from 55% to 85% over a couple months is a job where the make-ready process is starting to bite.
The real gold isn't the PPC number itself — it's the reasons for the misses. Every failed commitment gets a reason code: prerequisite not done, material late, weather, crew reassigned, design change, bad estimate. After a month you can see the pattern. If 40% of your misses trace to late material, you don't have a crew problem, you have a procurement problem, and no amount of yelling at foremen fixes procurement. Traditional variance analysis never surfaces that. PPC reason-tracking hands it to you on a plate.
Coordination: documents versus a room full of people
On a traditional job, coordination happens through paper. Here's the master schedule, here are the scopes, here's a monthly meeting to review status. Conflicts get discovered when two crews show up in the same corridor.
Last Planner coordinates through conversation. The weekly plan meeting puts the foremen in a room — or on a shared board — looking at each other's commitments for the same space and week. The plumber sees the drywaller is closing the wall Wednesday and says wait, I've got my top-out inspection Thursday, don't close it yet. That conflict just got solved in ninety seconds, standing up, instead of becoming a demo-and-rework ticket next week. You cannot get that from a document. You get it from people talking about the same near-term reality.
This is also where the two systems stop being enemies. The best-run jobs keep the CPM master schedule alive for what it's good at — milestones, long-lead procurement, contractual progress, and answering "what if" — while running the actual weekly execution on a look-ahead and weekly work plan. The look-ahead is the bridge: it takes the next slice of the master schedule and pressure-tests it against real constraints before anyone commits.
Where software actually helps (and where it doesn't)
You can run Last Planner with sticky notes on a wall, and plenty of great teams started exactly there. The physical board has real virtue — it's visible, it's tactile, everyone in the trailer sees it. What sticky notes can't do is travel, roll up, or remember. The plumber in the field can't see the wall. Nobody's tracking PPC trends across a hundred notes. Constraints written on a card get lost when the card falls off.
That's the honest case for software. A tool like LookAheadWall exists to make the look-ahead visual and location-based the way a good wall board is, but keep it live for the crew leader standing in the building on his phone, roll commitments up so you can actually see your PPC trend, and keep constraints from falling on the floor. The point isn't the software — it's that the discipline survives contact with a real jobsite and a distributed crew. If a tool makes the weekly plan harder to build or hides it behind training only a scheduler has, it's working against you. The whole appeal of Last Planner is that a foreman can participate without a CPM certification, because the concepts — is it ready, will you commit, did we deliver — are things every good foreman already thinks about.
Choosing what to run, and when
Keep leaning on traditional CPM scheduling when the question is strategic or contractual: your contract requires a submitted CPM schedule and monthly updates, the lender wants schedule-based progress, or you genuinely need logic-driven analysis to test alternatives on a large, complex sequence. That's real work CPM does well, and Last Planner won't replace it.
Lean hard on Last Planner and the weekly work plan when the problem is execution: multiple trades stacked in tight spaces, a job that keeps losing days to work that wasn't ready, or a field team you want actually engaged in planning instead of just receiving orders. If your crews are chronically standing around waiting on somebody else, that's not a CPM problem you can update your way out of. That's a make-ready problem, and make-ready is exactly what Last Planner is built to fix.
How to make the switch without blowing up the job
Nobody should rip out their master schedule and go all-in overnight. The transition that works is gradual and looks like this:
- Start with the weekly work plan. Before you touch anything else, get the foremen making real weekly commitments and start scoring PPC. Even alone, this exposes how much planned work was never actually ready.
- Add the look-ahead and constraint screening. Once weekly planning has traction, formalize the three-to-six-week look-ahead and start screening upcoming work for constraints with named owners and due dates.
- Work the reason codes. Track why commitments miss and attack the top one or two categories. This is where PPC turns from a scoreboard into an improvement engine.
- Keep the master schedule. Maintain CPM for milestones, procurement, and contractual reporting. Let the weekly reality feed progress back up to it, not the other way around.
Give it a couple of months. Early on your PPC will look ugly and your foremen will grumble about the extra meeting. That's normal — you're measuring reliability honestly for the first time, and honest numbers are uncomfortable. Stick with it and the meeting gets shorter, the constraints get cleared earlier, and the crews stop hitting dead work.
The bottom line
Traditional scheduling and the Last Planner System aren't a fight to win. One predicts the shape of the whole job so you can steer it at the milestone level. The other makes the next few weeks actually happen by clearing constraints, capturing commitments, and measuring whether you kept them. Run the master schedule for strategy and compliance, run the look-ahead and weekly plan for execution, and let the field reality flow back up. Do that, and the schedule on the trailer wall finally matches the job outside the window — which, after twenty years of them not matching, is the only thing that ever really mattered.