Ask a room full of superintendents whether the CPM schedule matters and you'll get a knowing smirk. On most jobs the master schedule lives in a binder or a PDF that got emailed at buyout, and the real work gets planned on a whiteboard Monday morning. Meanwhile the project controls group updates that same CPM schedule once a month, and the two documents drift so far apart that by the time you're framing the third floor they might as well be describing different buildings.
That drift is the whole problem. The critical path method and your near-term lookahead aren't rivals, and they aren't two ways of doing the same thing. They're two different altitudes looking at the same job, and a project runs well only when they stay tied together. Let me walk through what each one actually does, where each one lies to you, and how to make them feed each other instead of contradicting each other.
What CPM is actually good at
The critical path method builds the whole job as a network of activities linked by logic — this has to finish before that can start, these two can run in parallel, this one can slip a few days without hurting anything. Run the math across the network and two things fall out: the total project duration, and the critical path, which is the chain of activities where any delay pushes the end date. Everything off that path has float — slack you can spend before it becomes a problem.
That's genuinely valuable, and no lookahead replaces it. CPM answers questions a two-week whiteboard never can:
- Does the job still finish on time? If you slip the elevator installation four days, CPM tells you whether that eats float or moves substantial completion. Your lookahead has no idea.
- What's actually critical? On a busy floor, the crew that feels the most urgent is often sitting on two weeks of float, while the quiet long-lead item nobody's thinking about is the thing that will sink you.
- When do the long-lead items have to land? Switchgear, elevators, curtain wall, custom AHUs, generators — the stuff with 20-to-40-week lead times lives entirely outside any lookahead window. CPM is where you catch that the gear needed to be released three weeks ago.
If you've ever been in a delay claim, you already know CPM is also the language of schedule disputes. Float ownership, time-impact analysis, concurrent delay — those arguments all run on the CPM model. You want that model to be accurate and current, not a fossil.
Where CPM lies to you
Here's the catch, and every field guy has lived it. CPM assumes the world is logic-clean. It assumes that the moment activity A finishes, the crew, the material, the equipment, the inspection, and the clear work area for activity B are all magically present. It models finish-to-start relationships beautifully and models reality not at all.
The delays that actually kill your week almost never show up as CPM logic. They show up as: the light fixtures got stuck in customs, the fire-caulk sub is on another job until Thursday, the inspector can't come until you've got three floors ready so he'll batch them, the layout can't happen because the deck is still full of the drywall crew's stocked material. None of that is a dependency Primavera drew. All of it is why you didn't hit the plan.
The other quiet failure: a CPM schedule updated monthly is wrong the day after it's updated. Between updates it drifts, and everyone in the field knows it's drifting, so they stop trusting it, so they stop looking at it, so it drifts worse. A schedule nobody believes is worse than no schedule — it gives false comfort in the trailer while the field runs on tribal knowledge.
What the lookahead does that CPM can't
The lookahead — call it a three-, four-, or six-week window depending on your trade mix and lead times — exists to make near-term work ready. Its whole job is to catch the constraints CPM assumes away, far enough ahead that you can actually clear them.
The discipline that makes this work is the make-ready process: for every activity coming into the window, you ask whether it can actually be done, and you name what's missing. The common constraint buckets are worth memorizing because you'll check them a thousand times:
- Material — is it on site, or does it have a confirmed delivery date that lands before you need it?
- Information — is the RFI answered, the submittal approved, the detail resolved?
- Crew — is the sub committed with the right headcount, or are they promising bodies they don't have?
- Equipment — lift, crane pick, hoist window, welding machine?
- Predecessor work — is the area actually complete and inspected, not just "basically done"?
- Access and space — is the work area clear, safe, and not double-booked with another trade?
An activity doesn't earn a spot on next week's work plan until its constraints are cleared. That's the single most important rule in short-interval scheduling, and it's the one CPM structurally cannot enforce, because CPM doesn't know constraints exist. A tool like LookAheadWall earns its keep here by making those trade-flow sequences and constraints visual — you can see the sequence of hand-offs by location and spot the place where two crews are about to collide before they're standing in the same room arguing about it.
Two altitudes, one job
The clean way to think about it: CPM is the flight plan, the lookahead is the next few miles of road, and the weekly work plan is what your foot does on the pedal this week. You need all three, and they have to agree.
The connection runs both directions, and both directions are where most jobs break down.
Top down: every activity in your lookahead should trace back to a master-schedule activity. When you break a CPM line item like "Level 4 MEP rough-in" into the twelve real hand-offs it actually is — overhead plumbing, then overhead duct, then major electrical, then in-wall rough, coordinated by area — those pieces still have to add up to the CPM duration and hit the CPM milestone. If your detailed plan quietly needs three weeks for something the master schedule gave two, you want to know that in the trailer now, not on the floor later.
Bottom up: what actually happened in the field has to flow back and update the master schedule. This is the step everybody skips. Your lookahead knows the real percent-plan-complete, knows which crew fell behind, knows the fire-alarm rough started four days late. If that reality never reaches the CPM model, the master schedule keeps forecasting a finish date built on fiction. The most dangerous schedule on a job is a CPM that still says "on track" because nobody's been feeding it the truth.
The buffers CPM won't draw for you
CPM logic tends to butt activities end to end — predecessor finishes Friday, successor starts Monday. In the field that's how you get inspection failures and rework, because you left no room for the messy realities between trades. A good lookahead builds in the buffers a clean network diagram omits:
- Between frame and rough-in, give yourself a day or two for cleanup, punch of the framing, and the framing inspection — don't let the electrician start pulling in a wall the inspector hasn't signed.
- Before you close a wall, megger your homeruns and pressure-test the lines. Closing a wall over a bad run or an untested joint is the most expensive shortcut in the building.
- Build in the batch-inspection reality: many jurisdictions won't send an inspector for one floor, so plan your rough-in completions to hit inspectable chunks rather than trickling.
- Leave slack at the front of any activity that depends on a submittal or long-lead delivery — the CPM tie says "start Monday," but if the material clears the dock Monday you've already lost the week.
These aren't dependencies you'll find in the master schedule. They're the field judgment that turns a theoretically correct sequence into one that survives contact with an actual jobsite.
How the two levels fit the Last Planner mindset
If you've run pull planning, this all maps cleanly onto the Last Planner framework, and it's worth naming because it gives you a shared vocabulary with the trades. Phase pull-planning sits right under the CPM milestones — the trades themselves build the hand-off sequence for a phase, tied to the master dates. The lookahead is the make-ready engine that clears constraints in that phase. The weekly work plan is the commitment: only constraint-free work gets promised. And percent-plan-complete plus reasons-for-variance is the feedback loop that teaches you why you missed — which then corrects both your next lookahead and your CPM assumptions.
The number that ties it together is your PPC. If you're completing 50 percent of what you commit to each week, your problem usually isn't the crews — it's that you're committing to work that was never actually ready, which means your make-ready process is weak and your CPM durations were fantasy. Chasing the reasons-for-variance back up the chain is how the field educates the master schedule.
Who owns what — and the seam between them
On most jobs a scheduler or project controls person owns the CPM model in Primavera or MS Project, and the superintendent owns the lookahead and the weekly plan. That division is fine. The failure is when nobody owns the seam between them.
Somebody has to be responsible for the two-way flow: making sure lookahead activities trace to CPM lines, and making sure field actuals get back into the model before the monthly update. Make that an explicit job, put it on a recurring weekly touchpoint, and put both people in the same room looking at the same reality. The superintendent doesn't need to run a time-impact analysis, and the scheduler doesn't need to run your Monday huddle — but both need to understand how their level connects to the other, or the two documents drift right back apart.
The mistakes to watch for
- Two disconnected schedules. A pristine CPM in the trailer and a lively whiteboard on the floor that never talk to each other. Both can be excellent and the job still misses milestones.
- Great near-term, no horizon. Flawless lookahead execution that never lifts its eyes to the long-lead items and the critical path. You'll win every week and lose the job.
- Great CPM, no execution. A perfect network diagram that never gets translated into constraint-cleared weekly commitments is a theory paper, not a plan.
- Double entry. Typing the same activities into two systems by hand guarantees they'll disagree within a month. Wherever you can, let the tools link so the connection is structural, not a person's memory.
Bottom line
CPM and the lookahead aren't a choice you make; they're two instruments you fly at once. The master schedule tells you whether the whole job finishes on time and where the real risk lives. The lookahead turns that framework into work that's genuinely ready to build this week. Neither one is enough alone, and the crews that consistently deliver aren't the ones with the prettiest master schedule — they're the ones who keep the two altitudes honest with each other, week after week, so that the plan in the trailer and the work on the floor are describing the same building.