A three-week lookahead is only as good as the day you last touched it. I've walked onto plenty of jobs where the super proudly pulled up a lookahead that hadn't moved in a month — activities finished two weeks ago still sitting there, a "start" date that came and went, trades working off memory because nobody trusted the paper anymore. That's not a schedule. That's a fossil.
The lookahead earns its keep in the update, not the build. Building it once is a Tuesday afternoon. Keeping it honest week after week is the actual discipline, and it's the part that separates a plan crews follow from a plan they nod at and ignore. This is the routine I've settled on after a lot of years and a lot of jobs where I got it wrong first.
Pick your update window and defend it
The single biggest thing you can do is make the update happen at the same time every week, without fail. Crews plan their lives around rhythm. If your weekly work plan lands Thursday afternoon one week and Monday morning the next, foremen stop planning around it and start planning around what they saw last time you talked in the field.
My preference is to do the heavy update Thursday, so it's finished and distributed before the Friday coordination meeting, and everyone walks into Monday knowing the week. Some supers run it Monday morning off the weekend's completed work. Either works. What doesn't work is "whenever I get to it." Block ninety minutes on the calendar, treat it like a concrete pour, and don't let a change order or an owner walk-through eat it. The week you skip is the week the schedule quietly dies.
Walk the job before you touch the schedule
Do not update the lookahead from your trailer chair off what people told you Wednesday. Update it off what you can see. Before you sit down, walk the floors with the current week's plan in hand and physically confirm status: what's actually done, what's 80% and getting called done, what hasn't started that was supposed to.
The gap between reported progress and real progress is where schedules go to lie. Drywall "hung" that's missing the top three feet on two walls isn't hung — the taper can't start, and if you roll it forward as complete you've just baked a false start date into next week. Mark percent complete honestly, even when it's ugly. A schedule that admits you're behind is worth ten that pretend you're fine.
Clear out finished work, then pull the next week forward
Once you know real status, do the mechanical cleanup. Anything genuinely complete comes off the active board — it's done its job, and leaving it there just clutters the view and makes crews hunt for what's live. This is the housekeeping half of a rolling lookahead: as the front week closes out and drops off, a fresh week rolls in off the master schedule to keep the window three weeks deep.
Pulling that new week in is where you actually earn your paycheck, because it's not a copy-paste from the CPM. The master schedule tells you the intent; you translate it into location-based, crew-sized chunks that can be worked. "Level 3 interior rough-in, 3 weeks" on the master becomes, on your lookahead, "Level 3 east wing plumbing rough" and "Level 3 east wing electrical rough" broken by area and sequenced so two trades aren't fighting over the same ceiling on the same day. Software that lets you lay this out visually by area and connect the trade-flow sequence — LookAheadWall is built around exactly this kind of location-based planning — saves you from re-deriving the same coordination in your head every week. But the thinking is yours. The tool just holds it.
Run the constraints — this is the whole point
If you take one thing from this, take this: an activity does not belong on the plan until it's clear to work. The reason lookaheads fail isn't bad sequencing, it's activities that were never really ready getting scheduled anyway. Every week, go line by line through the incoming and current work and confirm the constraints are actually lifted, not "should be by then."
The constraint checklist I run on every new activity:
- Prerequisite work complete — the trade in front of you is actually done in that area, not "almost."
- Material on site — not "ordered," not "on the truck." On site, or with a delivery date you've personally confirmed with the vendor.
- Manpower committed — the sub has told you a crew size and a day, and you believe them.
- Information/RFI answered — no open questions that stop the work. An unanswered RFI on a detail is a hard constraint, not a footnote.
- Access and prior inspection — the area is reachable, and any required inspection on the preceding work has passed. You can't insulate over rough-in that hasn't been signed off.
- Equipment — lift, hoist time, pump, whatever the task needs is scheduled and not double-booked.
Anything with an open constraint is a candidate for the plan, not a commitment on it. Track those constraints and who owns clearing them, with a need-by date that's earlier than the work date. This is the backbone of the Last Planner approach, and it's the difference between a plan crews trust and a wish list. When your percent-plan-complete is low, the honest post-mortem almost always traces back to work that got scheduled with a constraint still hanging.
Adjust the sequence to the job you actually have
Field reality votes, and it doesn't care what the master schedule intended. Maybe the concrete guys got rained out and the whole east side slid two days. Maybe the mechanical sub surged three extra bodies and got ahead. Your update is where you resequence to the world as it is — reroute crews to areas that are actually ready, keep trades flowing instead of stacking, and protect the buffers that keep one slip from cascading.
On buffers: don't butt trades cheek to cheek. Frame-to-rough-in usually wants a day or two of daylight for cleanup, layout, and inspection before the next trade lands. Rough-in to cover (drywall, insulation, whatever closes the wall) needs the inspection sign-off in hand — schedule the inspection as its own line, because "we'll call for it" is how you lose two days waiting on the inspector. And before you ever close a wall, the electricians should megger their runs and the plumbers should hold pressure on the lines. Finding a bad run or a weeping joint after the rock is up is a demo-and-repair story nobody wants to tell the owner.
Pull the trade foremen in — don't just publish at them
The best update information doesn't live in your head; it lives with the foremen doing the work. Before you finalize, get their read. What are they committing to next week, honestly, given their manpower? What's going to bite us that I can't see from the floor? A drywall foreman knows his taper's three days out on another job before it ever shows up on your radar.
This is where the update stops being a document you hand down and becomes a plan the people building it actually made. Reliable-promise commitments — the foreman says "yes, that area, that day, that crew" and means it — are worth more than any date you assign top-down. Collect those, hold people to them, and track how often the commitment held. Over a few weeks the pattern tells you which subs' word you can build on and which ones you need to buffer around.
Sanity-check before you hit send
Before it goes out, run a two-minute logic pass. Is anything scheduled before its predecessor finishes? Any trade double-booked in the same area on the same day? Any activity still carrying an open constraint you forgot to clear? Are the durations sane — did somebody's optimistic one-day task quietly become a real three-day task? A schedule that shows two crews in the same corridor Wednesday isn't a plan, it's a fight you scheduled.
Catching that stuff before distribution costs you two minutes. Catching it after — when a sub has already mobilized to an area that isn't ready — costs you a crew day and a chunk of credibility. The whole reason to keep the plan clean is so that when you say "you're good to go," the trade believes you.
Distribute it like it matters
An updated schedule that lives on your laptop helped no one. The instant it's clean, push it to everyone who works off it — trade foremen, the subs' offices, the owner's rep, your own crew leads. Same channel every week so nobody's hunting for the current version. This is also where a shared app beats emailed PDFs: everyone opens the same live plan, the crew leaders can see their week on a phone in the field, and you never get the "which version am I looking at?" call. Version drift — three different printouts floating around, each a little wrong — is a quiet killer, and it's most of the argument for keeping one source of truth instead of a chain of attachments.
Keep the discipline, not the drama
None of this is complicated. It's a walk, a cleanup, a hard look at constraints, a conversation with your foremen, a logic check, and a send — the same six moves, same time, every week. The magic isn't in any one step; it's in doing it religiously so the plan is always current enough that people bet their week on it.
The supers whose jobs run smooth aren't the ones with the fanciest schedule. They're the ones whose lookahead was updated Thursday, every Thursday, and whose foremen know that when it says the area's ready, it's ready. Build that habit and the plan starts doing the work of keeping everyone honest — which is the entire point of having one.