Every Last Planner rollout I've seen eventually collides with the same question: are we actually getting more reliable, or does it just feel that way because we're having nicer meetings? The honest answer only comes from a handful of numbers, and most teams either track the wrong ones or track the right ones and then do nothing with them. This is a walk through the metrics that actually change how a job runs, what good and bad look like, and the traps that make the numbers lie to you.
The rule of thumb up front: measure the fewest things you'll actually act on. A dashboard with thirty gauges is a dashboard nobody reads.
PPC Is the Whole Ballgame — But It's Easy to Fake
Percent Plan Complete is the signature Last Planner metric, and for good reason. You take the commitments a crew made for the week, you count how many they actually finished as promised, and you divide. Ten tasks committed, seven done, PPC is 70 percent. That's it. No weighting by dollar value, no partial credit for "we started it." Done or not done.
The power of PPC is that it measures the one thing that wrecks jobsites: broken promises between trades. When the drywaller says he'll have three units hung by Thursday and he doesn't, the taper who staged his crew around that handoff eats the loss, and it ripples. PPC is the earliest, cheapest place to see that unreliability before it shows up as a blown milestone.
Here's the part the textbooks skip. PPC is trivially easy to game, and every team games it a little without meaning to. The most common trick is soft commitments — "we'll make good progress on the second floor." You can't fail a task that vague, so it always counts as complete and your PPC looks great while the job quietly slips. A real commitment is specific enough that anyone standing in the field can walk up and tell you yes or no: "Units 204 through 209, MEP rough-in complete and ready for inspection by Thursday." If you can't verify it from the hallway, it's not a commitment, it's a wish.
What the Number Should Actually Be
People obsess over the target. Rough guideposts, and treat them as guideposts, not grades:
- 60–70% — a team in its first month or two of real short-interval scheduling. Normal. Don't panic.
- 70–80% — you're finding your constraints before they bite. Good progress.
- 80–90% — a disciplined team with clean make-ready. This is a genuinely well-run look-ahead.
- Consistently 95%+ — be suspicious. This usually means people are sandbagging their commitments so they never miss, not that the job is a Swiss watch.
That last one trips up owners and executives who think higher is always better. A crew that commits to eight easy tasks and hits all eight looks better on paper than a crew that commits to fourteen aggressive ones and hits twelve — but the second crew moved far more work. The trend matters more than the absolute number. A team climbing from 62 to 78 over six weeks is healthier than one sitting flat at 88.
Slice PPC So It Tells You Where to Look
A single project-wide PPC is a smoke alarm — it tells you something's burning but not which room. Break it down and it starts naming names:
- By trade. When your overall PPC is 74 and the electrician sits at 91 while the fire-sprinkler sub grinds along at 48, you don't have a "project problem," you have a sprinkler problem. Go have that conversation.
- By area or floor. If levels 1–3 run clean and level 4 keeps missing, the issue is usually access, staging, or a stacked-trade conflict specific to that zone, not the crews.
- By phase. Rough-in coordination and finishes fail for completely different reasons. Knowing which phase drags tells you where to spend your planning energy next job.
This is the point where the arithmetic gets tedious by hand, and it's where look-ahead scheduling software earns its keep. When crews enter their weekly work plan and mark tasks complete in the same tool, PPC by trade, by area, and over time falls out automatically instead of living in a spreadsheet nobody updates after week three. That's the practical reason we built LookAheadWall the way we did — the number is worthless if producing it is a chore.
Variance: The Metric That Actually Fixes Things
PPC tells you that you missed. Variance tells you why, and why is where all the improvement lives. Every time a committed task doesn't get done, someone should capture the reason in plain language, then bucket it. Over a few weeks the buckets tell a story:
- Prerequisite work wasn't finished (the trade ahead of you slipped)
- Materials weren't on site or weren't the right ones
- Manpower — crew got pulled to another job or called in short
- Information — an RFI wasn't answered, a detail wasn't resolved
- Changed priorities — the super redirected the crew mid-week
- Weather, inspection delays, equipment
The magic is in the pattern. Any single miss is noise. But when you tally a month and find that 40 percent of your misses trace back to unanswered RFIs, you've stopped managing symptoms and found a root cause — your information flow is broken, and no amount of yelling at foremen will fix it. I've watched teams cut their misses nearly in half in a couple months not by working harder but by killing the one recurring variance category that was quietly bleeding them. If you capture variance and never analyze it, you've done the hard part and skipped the payoff.
Leading Metrics: See the Miss Before It Happens
PPC and variance are both lagging indicators — they report on a week that's already over. To actually steer, you want a couple of leading numbers that predict trouble before it lands:
Constraint status. As you pull work into your look-ahead window, each task has to clear its constraints — design answered, materials ordered and confirmed, prior trade done, equipment reserved, permit in hand. Count how many tasks two and three weeks out still have open constraints. That count is the single best predictor of next month's PPC. A wall of red constraints three weeks out is next month's blown week, and you're seeing it early enough to do something.
Workable backlog. This is the stack of tasks that are fully constraint-free and ready to go right now. A healthy backlog is a buffer: when a crew gets rained out of exterior work, the foreman pulls the next ready task instead of standing his people down. A thin or empty backlog means the first disruption idles a crew — you have no slack. Watch that backlog like a fuel gauge.
Constraint lead time. How far ahead you're identifying constraints before the work is scheduled. A team catching them six weeks out has room to chase a long-lead material or a slow RFI. A team catching them the Friday before is firefighting. This is the practical argument for a longer look-ahead window on jobs with long procurement — a two-week horizon simply can't surface a constraint that needs eight weeks to clear.
Make-Ready Health
The make-ready process — systematically screening upcoming work and clearing constraints so tasks become truly ready — is the engine underneath everything above. Two numbers keep it honest. Ready rate: of the tasks you planned to make ready this cycle, what fraction actually cleared all constraints on schedule? If that's low, your look-ahead is theater — you're listing work but not doing the legwork to enable it. And constraint resolution cycle time: once a constraint is flagged, how long until it's cleared? If constraints sit open for three weeks on average, tighten your two-week window and your make-ready is already too late to matter.
The Metrics That Waste Your Time
Just as important is knowing what to ignore. A few traps I've watched sink otherwise good rollouts:
- Over-measurement. Tracking fifteen metrics means nobody internalizes any of them. In the field, three or four numbers people can recite from memory beat a report nobody opens.
- Measurement without action. A PPC chart that gets presented every week and never changes a single decision is decoration. If the number never redirects a crew, reprioritizes a constraint, or triggers a hard conversation with a sub, stop producing it.
- Gaming and false precision. When PPC becomes a stick to beat trades with, foremen protect themselves by committing to less. You optimize the metric and starve the job. Keep PPC a learning tool, not a scoreboard for punishment — the moment it becomes the latter, the data goes soft and stops meaning anything.
- Portfolio comparison abuse. Ranking two projects by raw PPC ignores that a tricky hospital renovation and a repetitive garden-apartment build are not the same animal. Compare a project to its own trend, not to a job that isn't its peer.
A Review Rhythm That Sticks
Match the metric to the cadence or it won't get looked at:
- Weekly, in the plan meeting. Last week's PPC, the top variance reasons, and the constraint status for the next two to three weeks. This is where the numbers actually change decisions — keep it to ten minutes, keep it in front of the people who made the commitments.
- Monthly. Variance patterns and PPC trend by trade and area. This is where root causes surface and process changes get made.
- End of project. The honest retrospective — what your reliability curve looked like, which trades and phases fought you, and the two or three lessons worth carrying to the next job.
Weekly PPC belongs to the field team and the subs who made the commitments. Monthly trends belong to the PM. Portfolio views belong to the office. Same underlying data, different altitude — and if you're running the look-ahead in software rather than on butcher paper, every layer draws from one source instead of three people maintaining three versions of the truth.
The Short Version
If you remember nothing else: PPC tells you whether you're reliable, variance tells you why you're not, and open constraints tell you what's about to go wrong next. Everything else is supporting detail. Insist on commitments specific enough to verify from the hallway, capture every miss with a real reason, and actually act on the one or two patterns that keep showing up. A team that does only that — and does it every single week without fail — will run circles around one drowning in dashboards it never reads. The data is the easy part. The discipline to use it is the whole job.