Every superintendent already ranks their subs in their head. You know which drywall crew shows up Monday with the right manpower and which one you have to call three times before they send two guys and a helper. The problem with keeping that scorecard in your head is that it walks out the door when you retire, it argues with the estimator who only sees the low number, and it can't tell you whether a trade is actually getting worse or you just had one bad month with them. Performance tracking software exists to get that judgment out of your head and into something the whole company can see, defend, and act on.
The catch is that most of what gets sold as "subcontractor management" tracks the easy stuff — certificates of insurance, signed contracts, whether the pay app came in on time. That matters for the office, but it tells you nothing about whether a sub can hold a schedule. The performance data that actually protects your job comes off the field, and most of it comes straight out of your weekly planning if you're capturing it. Here's what's worth tracking, how to measure it without lying to yourself, and where the numbers quietly mislead you.
Start with schedule reliability, because it drives everything else
If you track one thing, track whether a sub does what they committed to do this week. This is the heart of short-interval scheduling and it's the most honest performance signal you have, because it's measured against a promise the sub made themselves, not against a plan someone handed them.
The metric borrowed from Last Planner is Percent Plan Complete (PPC): of the tasks a trade committed to in the weekly work plan, what fraction were actually done and done complete by Friday? Count a task as complete only if it's truly finished — no "90% done, we'll knock out the rest Monday." Partial credit is how you fool yourself. A crew running 85–90% PPC week over week is a partner you can build a look-ahead around. A crew at 50% is telling you their commitments are guesses, and every downstream trade you sequenced behind them is now exposed.
The more useful half of PPC is the reasons for the misses. When a task doesn't get done, log why in a plain category: no manpower, waiting on the trade ahead, material not on site, RFI unanswered, weather, rework. After a month those reason codes tell you the real story. A framer at 60% PPC because they're chronically short-handed is a different problem than a framer at 60% because the concrete crew keeps handing them slabs late. One is a sub performance issue; the other is a sequencing issue you own. Software that captures commitments and variance reasons inside the weekly plan — which is exactly what a tool like LookAheadWall is built to do — turns that into a running record instead of a memory.
Quality: measure rework, not just punch counts
Quality is where raw counts lie to you. A big mechanical sub will always rack up more punch items than the guy who hangs one bathroom accessory, simply because they touched more of the building. Normalize it. Punch items per unit of work — per floor, per unit, per thousand square feet — is comparable across trades and across jobs. Raw totals aren't.
The metric that actually costs you money is rework: work that passed once and had to be torn out or redone. Track first-time inspection pass rate (did the rough-in pass on the first walk, or did the AHJ red-tag it?) and track callbacks. A tile setter with a clean punch but a habit of getting kicked back on the first inspection is costing you float you can't see in the punch numbers, because every failed inspection stalls close-in for everyone behind them. Watch for patterns, too — if the same deficiency shows up on three jobs (say, a plumber who never quite gets the escutcheons right), that's a training conversation, not a punch item.
Safety: separate leading indicators from the EMR
Everyone quotes the Experience Modification Rate because it's on the prequal form, but EMR is a lagging, insurance-driven number that trails real behavior by years. It belongs in selection, not in day-to-day management. For managing a sub on your active job, watch the leading indicators: observed violations per week, whether their competent person actually shows up to the safety meeting, near-miss reporting (a crew that reports near-misses is usually the safer one, counterintuitive as that sounds), and training/certification currency for the work in front of them. A trade with a beautiful EMR and three fall-protection violations this month is a claim waiting to happen, and the EMR won't warn you until it's too late.
Communication and administrative metrics — useful, but keep them in their lane
Responsiveness is trackable and it matters: RFI turnaround, submittal timeliness, whether the foreman answers the phone when you've got a change in front of the crew. Administrative items — insurance currency, pay-app accuracy, closeout document completeness — matter to the office and to your risk exposure. Track them. But don't let a sub with slow email drag down a scorecard when their crew is the best on the job. Weight these categories honestly against what actually affects the schedule. A great installer with a mediocre office is still a great installer.
Rolling it up: scorecards that don't lie
A scorecard consolidates schedule, quality, safety, communication, and administrative performance into one view per trade. It's genuinely useful — and it's genuinely easy to build one that's worse than useless. A few rules that keep it honest:
- Weight to your reality. Schedule reliability and rework should dominate. If your weighting lets a perfect COI compensate for a crew that blows every commitment, you've built a compliance report, not a performance tool.
- Never rank on a single job. One bad project — bad plans, a brutal owner, a design that changed weekly — can tank a good sub. Require a minimum number of data points before a number means anything.
- Keep the context attached. A 55% PPC on a job where you fed the trade late is a mark against you, not them. If your reason codes don't travel with the score, the score is an accusation with no defense.
- Show the trend, not just the snapshot. A sub climbing from 60% to 85% over a year is a keeper you're actively developing. A sub sliding the other way is a warning. The direction matters more than the number.
The failure mode nobody warns you about: gaming and grudges
The day performance scores start driving award decisions is the day people start managing the score instead of the work. Foremen learn to commit to less in the weekly plan so their PPC looks better — sandbagging. The fix is to watch commitment volume alongside completion rate; a crew that commits to two tasks and finishes both isn't outperforming the crew that commits to ten and finishes eight.
The uglier failure is the grudge. A super who got burned by a sub in 2019 can quietly bury them in the numbers for years. Objective, field-captured metrics with the reasons attached are the antidote — the point of writing it down is to force the data to argue back when your gut is wrong. And give the sub a way to see and contest their number. A trade that can look at their own scorecard and say "that late start was your slab, not my crew" is a partner. A trade that finds out they've been blackballed on numbers they never saw is an adversary, and word travels fast in a bid market that's tighter than any of us would like.
Closing the loop: the data is worthless if it dies in a spreadsheet
Here's the part most firms skip. You track all this, build the scorecards, and then it lives in a file the estimator never opens at bid time and the sub never hears about. Performance tracking only pays off in two places: at selection, where a proven schedule performer earns preferred status and a shaky one gets a hard look, and in the field, where you sit down with a foreman mid-job, show them their PPC and their miss reasons, and fix the actual problem before it eats another three weeks of float.
Do it as a conversation, not a report card. "Your crew's been at 60% and every miss is 'no manpower' — what's going on, and what do you need from me?" gets you further than a red cell on a dashboard ever will. Half the time the honest answer is something you can fix: you were feeding them late, or the submittal was stuck in your own office. That's the quiet value of tracking this well — it doesn't just grade the subs, it holds up a mirror to how the job is being run. The numbers that come out of a disciplined weekly work plan are the same numbers that make you a better superintendent, and that's worth more than any scorecard you'll ever hand upstairs.