On a simple job, a bad schedule embarrasses you. On a complex one, it buries you. I learned that the hard way on a mid-rise where the CPM schedule said we were fine right up until the week three floors of drywall stopped dead because nobody had confirmed the fire-caulk inspection would release the shafts. The Gantt chart was green. The building wasn't. That gap between what the schedule says and what's actually happening is exactly where complex projects go to die, and it's the single best reason the Last Planner System, backed by decent software, earns its keep on hard jobs.
The short version: complexity doesn't break the Last Planner System, it's the reason you use it. Let me walk through why, with the parts that actually matter on site.
What "complex" really means on a jobsite
People throw "complex" around like it just means "big." It doesn't. A 400-unit garden apartment complex is big but often simple — the same twelve activities repeat 400 times. A 90,000 SF hospital fit-out is not big by square foot but it will humble you, because complexity is about how tightly the pieces are coupled, not how many there are.
The dimensions that actually hurt:
- Trade density in a small footprint. A mechanical room where pipefitters, sheet metal, electricians, controls, and insulators all need the same 200 square feet in the same two weeks.
- Long constraint lead times. Switchgear at 30+ weeks, air handlers at 20, a single owner finish decision that gates three trades and takes a month to land.
- Sequential dependencies with inspection gates. Rough-in, then inspection, then insulation, then close, then a second inspection. Every gate is a place the whole train can stall.
- Low float. Compressed schedules where there's no slack to absorb a two-day miss, so every miss cascades.
When those stack up, the traditional master schedule — built once, months out, by someone in an office — starts lying to you. Not out of malice. It's just that the assumptions baked in on day one don't survive contact with the field.
Why the master schedule alone can't carry a hard job
A CPM master schedule is a prediction. On a simple job the prediction mostly holds because there aren't many ways for it to be wrong. On a complex job, the prediction is a tower of assumptions — this trade will be staffed, that submittal will be approved, this inspector will show, that deck will be dry — and every assumption is a coin flip. Flip 2,000 coins and some come up tails. The math is against you.
The master schedule is still valuable. You need it to hold the milestones, the phasing logic, and the contractual dates. What it can't do is tell your foreman what to actually build next Tuesday, because by the time it's updated, the field has already moved on. That's the job of the look-ahead and the weekly work plan, and that's where Last Planner lives.
Push the master date down into a look-ahead you can actually work
The engine of the whole system is the rolling look-ahead — usually a six-week window, sometimes three or four on a fast job. You take the milestones the master schedule demands and you break them down into the real, buildable tasks that get you there, then you screen every one of those tasks for what's stopping it.
The look-ahead's real job isn't to schedule work. It's to make work ready. A task doesn't get to enter the weekly plan until its constraints are cleared: material on site, prerequisite work complete and inspected, information in hand (approved shop drawings, an answered RFI), crew available, equipment lined up, and the area physically accessible. If any of those is open, the task stays in the look-ahead as a constraint to chase, not a commitment to make.
On a complex job this screening is the whole ballgame. Constraints don't show up on a schedule; they show up when a foreman looks three weeks out and says "I can't start the overhead rough until the structural steel connections get their final bolt-up sign-off, and that's on nobody's list." That sentence, said out loud in a weekly meeting five weeks before the work, is worth more than any Gantt chart. Good look-ahead software matters here mainly because it keeps that constraint from evaporating — it gets an owner, a need-by date, and a status that somebody has to answer for every week.
Distribute the planning — no one person can hold a complex job in their head
Here's the thing nobody wants to admit: on a genuinely complex project, not one person alive understands the whole thing well enough to sequence it correctly from an office. The superintendent doesn't know the fitter's rigging plan. The PM doesn't know why the drywall foreman wants to hang the corridors before the rooms. That knowledge lives in the trade foremen's heads, and the Last Planner System is built to pull it out.
That's the quiet genius of it. The people who make the commitments are the last planners — the foremen who will actually direct the crews. They plan their own week, in a room, out loud, in front of the trades they hand off to and receive from. When the framer says "I'll have the north corridor topped out Thursday," the drywall foreman in the same room can say "then I'm starting there Friday, so I need my board loaded Wednesday." That handoff got negotiated by the two people who own it, not dictated by a chart. A shared weekly work plan that every foreman can see — on a wall, or in an app the crew leaders carry — is what keeps those commitments visible instead of buried in someone's notebook.
The trade-coordination gotchas that actually sink complex jobs
After enough years you stop being surprised by the same failures. These are the ones that eat complex projects, and how the planning process catches them:
- The MEP overhead scrum. In a tight ceiling, sequence is everything: big duct first, then plumbing waste (it needs its slope), then hydronic and domestic, then conduit and cable tray (electricians can dodge around obstacles better than a fitter can). Get this order wrong and you're relocating installed work. In the weekly plan this shows up as who claims the ceiling grid for which days — one trade at a time, not four at once.
- The inspection gap. Frame-to-rough-in wants a 1–2 day buffer for cleanup and inspection before anyone insulates or closes a wall. Don't schedule cover-up the same day as the inspection. If the inspector reds it, you're opening walls back up, and on a complex job that ripples for weeks.
- Wet trades before dry. If the fireproofing or the topping slab isn't scheduled and cured before finishes, you're either waiting or ruining work. Cure times are constraints — put a real date on them.
- The "we'll figure out access later" trap. Two trades needing the same lift, the same hoist window, or the same laydown at once. This never resolves itself. It resolves in the weekly meeting or it resolves in a fistfight in the parking lot.
- Silent information constraints. An unanswered RFI or an unapproved submittal is a constraint even though it never touches a tool. Link it to the task it gates and chase it on the look-ahead, or you'll find out about it the morning the crew shows up with nothing to build.
Notice none of these are exotic. They're ordinary. Complexity just means several of them are live at once, and the weekly plan is the only place they collide early enough to fix cheaply.
Scale the process, don't abandon it
When a project gets big enough that one weekly meeting can't hold every trade, don't give up on the discipline — tier it. Run area-level or phase-level planning (the podium, the tower floors, the central plant) with the trades active in that zone, then a shorter project-level coordination session that stitches the areas together and handles the cross-zone handoffs — the tower crane, the temp power, the shared hoist. The trap is letting the areas plan in isolation until their work collides at a shared resource. The integration meeting exists specifically to catch that.
Expect lower PPC early, and read the trend not the number
Percent Plan Complete — the share of weekly commitments you actually finished — is your honesty meter. Complex projects run lower PPC than simple ones, especially at the start, and that's normal. More constraints per task, more handoffs, more ways to miss. A team new to the process might start in the 50s. That's fine. What matters is the trajectory and, more than the number, the reasons for variance.
Track why commitments failed — "prerequisite not complete," "material late," "changed priority," "no info," "crew short" — and after a few weeks the pattern names your real problem. If half your misses are "prerequisite not complete," your trades are over-promising and your handoffs are the weak link. If they're "no info," your RFI and submittal pipeline is the bottleneck, not the field. That's the learning loop, and it's worth more than any single week's percentage. A team that drives PPC from 55 to 80 over a project has learned to plan; a team stuck at a flat number is filling in a form.
Start early, before complexity peaks
The best time to stand up this process is before the job gets hard — during site work and structure, when there's slack and the stakes are lower. You want the weekly-planning habit, the trade relationships, and the constraint-tracking discipline already in place by the time the interior fit-out turns into a knife fight for space. Trying to install a new planning culture in the middle of the crunch, when everyone's already stressed, almost never takes. Get the reps in while it's easy so the muscle memory is there when it counts.
The bottom line
Complex projects fail for a specific, predictable reason: too many tightly-coupled commitments, made too far in advance, by people too far from the work, with no reliable way to catch the constraints before they detonate. The Last Planner System attacks exactly that — it makes work ready before it's committed, puts the commitment in the hands of the foreman who owns it, and surfaces constraints while there's still time to clear them. Software like LookAheadWall earns its place by making the weekly plan and the constraint log visible to everyone at once and putting the current week in the crew leaders' hands, but the discipline is the thing that saves the job. The tool just keeps you honest about it.
Do the boring work every week — screen the constraints, make the handoffs out loud, track why you missed — and the complexity stops being the thing that buries you. It becomes just another set of problems you saw coming.