I've watched the Last Planner System get rolled out on jobs three different ways. Once it stuck and the job finished early. Once it limped along as a Tuesday meeting nobody prepared for. And once it died in six weeks because the super treated the software like a stick to beat foremen with. The software wasn't the difference in any of those. The difference was whether the crews believed the plan was theirs.
So before we talk about how to stand up a tool, let me be blunt about what you're actually buying. Last Planner isn't scheduling software. It's a way of running the last few weeks of planning collaboratively, so that the guys doing the work make the promises, constraints get cleared before work starts, and you measure how often the plan held up. The software just carries the paperwork. If your planning is broken, a slicker tool makes it broken faster.
Here's how to actually get it running on a live job without it collapsing.
Start on one job, not the whole company
Pick a pilot. Resist the urge to roll it out portfolio-wide because a VP got excited at a lean conference. You want one job with a super who's genuinely curious — not the one who's been doing it "his way" for thirty years and resents the idea, and not the young guy who'll agree to anything and follow through on nothing. Curious and stubborn is the sweet spot.
The job itself should have enough runway to see change. Anything under four months and you'll barely climb the learning curve before you're chasing punch list. Six months and up is where the habit has room to form. And it should have real trade coordination pressure — a straightforward tenant improvement with two subs on site doesn't stress the system enough to prove anything. You want overlapping trades stacking in the same areas. That's where reliable weekly planning earns its keep.
One more thing about the pilot: tell people it's a pilot. Say out loud that you expect the first few weeks to be rough and that you're not grading anyone yet. Crews smell a trap. Take the trap off the table early.
Get the super and PM bought in before anyone touches the tool
This is the step everyone shortcuts, and it's the one that decides the outcome. If the superintendent doesn't believe collaborative planning beats his gut and his whiteboard, no software on earth will save it. He runs the trailer. The crews take their cues from him. If he shows up late to the planning session, skips it when it's busy, or overrides commitments on a whim, everyone learns the meeting is theater.
What buy-in actually means in practice:
- The weekly session is sacred. It happens at the same time every week and the super is in the room, engaged, every single time — especially when the job is on fire. That's exactly when the crews are watching to see if this is real.
- The PM protects the time and cares about the results. If Percent Plan Complete never gets discussed, foremen learn it's a number nobody reads.
- Leadership doesn't panic when the first PPC comes in at 55%. That number is supposed to start low. A low honest number is worth more than a high fake one.
If you can't get the super genuinely on board, pick a different super or a different job. I mean that literally. It's cheaper to change the pilot than to burn the crews' patience on a rollout the field leader doesn't own.
Configure the tool to match the job — then stop
When you set up the software, the temptation is to model everything on day one: every phase, every sub-area, a dozen constraint categories, custom fields for fields you'll never fill. Don't. Over-configuration is the quiet killer. The crews open a screen that looks like a tax form and check out.
Configure the minimum that reflects how the job is actually broken up:
- Areas and phases that match how the field talks. If the crews call it "Level 3 east" and "the podium," name it that in the tool. Don't invent a coding scheme they have to translate in their heads.
- A short, honest list of constraint types. Materials, information (RFIs and submittals), predecessor work, labor, equipment, permit/inspection. Six or seven buckets covers 95% of what actually stops work. You can add more later if a real pattern demands it.
- A look-ahead window that fits the trade with the longest lead time on the job. Three weeks is the floor. If you've got long-lead material or a permit-and-inspection gauntlet, run a six-week look-ahead so constraints surface with time to clear them. The window exists to buy lead time, not to look impressive.
Set up accounts for the people who'll actually touch it — super, foremen, key sub leads — and leave it there. You can always add structure. You can rarely claw back the goodwill you lose making it complicated on day one. This is the same reason a good look-ahead scheduling tool like LookAheadWall defaults to a clean visual board instead of a spreadsheet with forty columns: the crew has to want to look at it.
Run the first session so it feels like planning, not a status report
The first weekly work plan meeting sets the tone for the whole job. The single most important thing you do in that room is get the language right. You are not assigning work. You are asking each trade leader what they can promise to finish this week. Those are different things, and crews know the difference instantly.
A rhythm that works:
- Walk the look-ahead. Go activity by activity through the next few weeks and ask, out loud, "what's stopping this?" Every constraint gets named, gets an owner, and gets a date. No owner, no date, it's not managed — it's a wish.
- Pull only clean work into this week. An activity earns its way into the weekly plan when its constraints are actually cleared — material on site, area handed off, inspection passed, RFI answered. This is the make-ready discipline, and it's the whole ballgame. Planning work you know is blocked is how you manufacture a low PPC and a demoralized crew.
- Get the promises. Ask each foreman what they'll deliver — not what they hope for, what they'll commit to. Then check the handoffs. If the electricians need rooms 210 through 216 by Wednesday, does the framer have those rooms ready by Tuesday end of day? That one question, asked every week, prevents more delay than any Gantt chart.
- Write it down where everyone sees it. Capture the commitments in the tool during the meeting, on the screen, so the plan is a shared record and not one guy's notes.
End with a plan every person in the room can repeat back. If they can't, you don't have commitments — you have attendance.
Through the week: track light, don't hover
Between sessions, your job is to watch for the plan cracking, not to stand over crews with a clipboard. When a promised task slips or a new constraint pops up, capture it while it's fresh — you'll need it Friday. Update the look-ahead so it reflects what's actually happening, not the pretty version from Monday. A rolling plan that lies to you is worse than no plan.
The mobile side matters here more than people expect. A crew leader who can pull the week's plan up on his phone at the gang box, and mark work done from where he's standing, keeps the board honest without a nightly trip to the trailer. The moment updating the plan becomes a chore that happens once a week under duress, the data goes stale and the whole thing turns into fiction.
Review before you plan — this is where the value hides
Every session, before you plan the coming week, you spend fifteen or twenty minutes on last week. This is the part teams skip when they're busy, and skipping it guts the entire system. Two numbers and one conversation:
- PPC. Of the tasks you promised last week, what percent actually finished — done-done, not "80% there"? A partial is a miss. The tool should total this automatically so nobody argues the math.
- The reasons for the misses. For every task that didn't complete, categorize why. Materials late? Prior trade not done? RFI unanswered? Crew pulled to another job? This is the gold. One miss is noise. The same category showing up three weeks running is a systemic problem you can now go fix.
Watch for patterns and act on them. If "predecessor not complete" keeps killing your week, your handoffs are the problem and your make-ready is too loose. If "materials" dominates, procurement isn't tracking to the look-ahead. The whole point of measuring PPC isn't the score — it's the reasons behind the score. A team that religiously runs the miss analysis and does nothing with it is just keeping a diary.
What actually goes wrong — and the numbers to expect
The failure modes are boringly consistent across every job I've seen. Learn them so you can catch yourself:
- The super dictates the commitments. The second a foreman is told what to promise instead of asked, it stops being Last Planner and becomes the same top-down schedule in a new outfit. The promises have to be theirs or they won't own the misses.
- The session gets cancelled when things get busy. Cancel it twice and it's dead. The busy weeks are precisely when the coordination pays off most.
- Gaming the number. Foremen figure out that promising less guarantees a high PPC, so they sandbag. If your PPC is a flat 98% and the job is still behind, someone's under-committing. A healthy PPC has some misses in it — that means people are stretching honestly.
- Blocked work in the weekly plan. Pulling activities you know aren't ready, then acting surprised when they don't finish. Make-ready discipline exists precisely to stop this.
On the numbers: expect PPC to start somewhere around 50 to 60% on a real job with honest reporting. That's normal. With disciplined weekly sessions and real miss analysis, you climb into the 80s over a couple of months. Don't chase 100 — a team pinned at 100 is a team that stopped promising anything hard.
A realistic timeline, if you want one to set expectations with leadership: a week or two to align the super and PM and set up the tool, another couple to train hands-on with the actual job's data and run the first sessions, then a solid month of just building the weekly habit before you tighten anything. Constraint management and pulling more subs into the room comes after the base rhythm is boring and reliable. Integrating with the master CPM schedule and cost — that's a "once it's working" problem, not a launch problem.
Then, and only then, standardize
When the pilot works — and you'll know, because the crews start policing their own handoffs without you asking — that's when you write down how you run it and take it to the next job. Name a champion or two who've lived it to help the next super over the hump. Set a portfolio PPC target if you must, but keep it honest.
The thing to remember through all of it: the tool doesn't implement Last Planner. People do. Good look-ahead scheduling software removes the friction — it keeps the plan visible, does the PPC math, puts the week in a crew leader's pocket, and stops the meeting from turning into a spreadsheet-wrestling match. That's real value. But it's leverage on a discipline you already have to build. Build the discipline first, keep the tool simple, ask for promises instead of handing out orders, and review your misses like they matter. Do that and the software earns its keep. Skip it and you've just bought a very organized way to be late.