Every super I know has a gut feel for whether the job is running well. You can walk a floor and smell trouble — trades stacked on top of each other, material in the wrong bay, a foreman who won't look you in the eye at the standup. The problem with gut feel is that it doesn't survive a meeting with the owner. When they ask why the drywall date slipped, "it just felt tight" doesn't hold up. A handful of numbers, tracked honestly week over week, does. This is about which numbers actually earn their keep, how to read them, and how to keep them from lying to you.
You don't need a data science degree or a wall of dashboards. You need maybe five metrics you trust and look at every Friday. The rest is noise until you've mastered those.
Start With Percent Plan Complete
Percent Plan Complete (PPC) is the one metric to start with, and honestly the one most crews never outgrow. The math is dead simple: of the tasks a crew committed to in this week's plan, how many actually got done? Ten committed, seven done, that's 70% PPC. Count a task complete only if it's 100% finished the way it was promised — not "mostly," not "we'll knock out the punch Monday." Partial credit is how you lie to yourself.
The number itself matters less than the trend. A single week of 65% tells you almost nothing; six weeks climbing from 55% to 78% tells you the crew is learning to make promises they can keep. That's the whole game. PPC isn't a productivity score — a crew can hang plenty of drywall and still blow their commitments because they committed to the wrong things. It measures the reliability of the plan, which is the thing that actually lets the next trade start on time.
What's a Good Number, Really
New teams almost always start in the 50–60% range, and that's normal. Foremen aren't used to committing to specific work in a specific week, so they over-promise. Anything in the high 70s to low 80s consistently means the planning discipline has taken hold. Chasing 100% is a mistake — a crew reporting 100% every week is either sandbagging their commitments or lying, and I've never seen a third option.
Don't wave the target around as a stick. The first time you turn PPC into a performance review number, foremen stop committing to anything they aren't dead certain of, and your plan gets so conservative it's useless. The point of the number is to find out where the plan broke, not to grade the man who reported it.
The Reasons Behind the Misses Are the Real Gold
PPC tells you a task didn't finish. The reason it didn't finish is where the money is. Every time a committed task misses, write down why in a single category. Keep the list short and consistent so you can count them:
- Prerequisite work — the trade ahead of them wasn't done (the number-one killer on most jobs)
- Materials — not on site, wrong material, or buried behind something else
- Manpower — crew got pulled to another job or showed short
- Equipment — lift down, no power, tools not staged
- Information — an RFI still open, missing shop drawing, unclear detail
- Directive change — owner or design change blew up the plan
- Weather — the honest one, and the one people hide behind
- Rework — had to tear out and redo
One miss is a bad day. Five weeks where "prerequisite work" is your top reason is a systemic problem — your trade flow is out of sequence, and no amount of yelling at the drywall foreman fixes it. That's a scheduling fix, not a manpower fix. When you tally these reasons over a month, the pattern practically writes your improvement plan for you. On one mid-rise I ran, "materials" quietly topped the list for six straight weeks until we finally traced it to a single distracted PM who'd stopped releasing submittals. The number found him; the walk-through never would have.
Measure Whether Work Is Actually Ready
Here's the metric that separates crews who firefight from crews who plan: Tasks Made Ready. Before a task ever lands in the weekly work plan, it should clear the make-ready process — all its constraints knocked down. TMR asks: of the tasks you pulled into this week's plan, how many were genuinely constraint-free when they got there?
If your PPC is low, TMR usually tells you why. You can't complete work that was never ready to start. A crew committing to tasks that still have open constraints is committing to hope. The fix isn't to push the crew harder in the weekly meeting — it's to push the make-ready work further upstream into the look-ahead, where you've got two, three, six weeks to clear constraints before the work goes live.
That's the real point of a rolling look-ahead: it's a constraint-removal machine, not a prettier bar chart. The weekly plan should only ever contain work that's been screened clean. Tools like LookAheadWall earn their place here because the look-ahead and the weekly plan live in the same view — you can see a task marching toward its start week with a constraint still hanging on it, and go kill that constraint before it kills your PPC.
Track Constraints Coming In and Going Out
Tie a couple of simple counts to your make-ready process. How many constraints did the team identify in the look-ahead this week? How many got resolved before their task hit the plan? And the one that stings — how many constraints got discovered in the weekly meeting that should've been caught two weeks earlier?
That last count is a leading indicator of pain. Constraints discovered late are the ones that blow up your week, because there's no runway left to fix them. If you're consistently finding constraints in the weekly standup, your look-ahead horizon is too short or nobody's actually screening it. A healthy operation catches the overwhelming majority of constraints weeks out and walks into the weekly meeting with clean, ready work.
Break PPC Down by Trade
A project-wide PPC of 72% hides a lot. Split it by trade and the picture sharpens fast. Maybe the framers and electricians are cruising at 85% and the whole number is getting dragged down by one mechanical sub sitting at 45%. Now you know exactly where to spend your Monday.
By-trade PPC also keeps you honest about who owns a problem. A sub sitting at 45% might genuinely be short on manpower — or they might be starved because they're always waiting on the trade ahead of them, in which case the low number is really a sequencing failure that shows up on their scorecard. Cross it against your variance reasons before you go chew somebody out. The number points you at the room; you still have to walk in and look.
Leading vs. Lagging — Know the Difference
PPC is a lagging indicator. It tells you what already happened last week; you can't change it. That's fine, but you can't steer with the rearview mirror alone. The leading indicators are the ones that predict next week before it arrives:
- Constraint status in your look-ahead — how much of the work slated for two weeks out is already clean
- TMR heading into the plan — are you committing to ready work or hopeful work
- Late-discovered constraints — trending up means next week's PPC is about to drop
Watch the leading numbers to steer, and use the lagging numbers to confirm whether your steering worked. A crew whose look-ahead is full of clean, ready work will post good PPC — you'll practically be able to call it a week out. That predictive quality is the entire reason short-interval scheduling beats a static Gantt chart: the Gantt tells you where you were supposed to be, the metrics tell you where you're actually headed.
A Few Honest Traps
Metrics get gamed the moment they get weaponized, so name the traps out loud:
Sandbagging. If a foreman learns that high PPC earns praise and low PPC earns a lecture, he'll commit to only the easy, certain tasks and leave the risky work off the plan entirely. PPC hits 95% and the job still slips, because the hard work is happening off the books. Watch for a suspiciously high PPC paired with a schedule that's quietly sliding.
Redefining "done." The instant "done" starts meaning "80% done," the number is worthless. Hold the line on complete-means-complete, punch and all.
Tunnel vision on one number. No single metric survives contact with a determined optimizer. PPC alongside variance reasons alongside TMR is much harder to fake than PPC alone — to cheat all three at once, a crew would basically have to do the actual work.
Keep the Cadence Simple
Calculate PPC and log your variance reasons every single week, right after the weekly plan closes — it takes ten minutes and the memory is fresh. Once a month, step back and read the trends: is PPC climbing, which variance reason keeps topping the list, is TMR improving. At project close, write down the story the numbers told so the next job starts smarter than this one did.
Don't build a metrics program so heavy that nobody keeps it up. Five numbers you actually look at beat twenty you collect and ignore. Good look-ahead scheduling software will crunch PPC and roll up variance reasons for you so the tracking isn't another job on your plate — but the discipline matters more than the tool. A superintendent with a whiteboard, an honest count, and the guts to write down why work missed will run circles around a crew with a fancy dashboard nobody trusts.
Start with PPC and variance reasons. Add TMR when those feel like second nature. Read the trends, chase the patterns, and fix the process — not the person. Do that for a few months and you'll stop guessing whether the job is healthy. You'll know, and so will the owner.