Walk into most trailers and you'll find a dashboard bolted onto a big TV that nobody looks at after the first week. It's got a pie chart of the budget, a gauge that says the job is "72% complete," and a scrolling list of open RFIs. It's colorful. It's also useless, because none of it tells the super what to do before lunch. A good construction dashboard isn't a status museum you visit once a month. It's a working surface you glance at every morning to decide where the crews go and what's about to bite you.
I've watched a lot of software promise "visibility" and deliver a wall of numbers. The teams that actually run tight schedules use dashboards the way you'd use the wall of a look-ahead: to see what's in front of you this week, spot the constraint before it becomes a stop-work, and move. Below is what separates a dashboard that changes behavior from one that just decorates the trailer.
Start From the Question, Not the Metric
The failure mode of almost every construction dashboard is that it displays what's easy to compute instead of what somebody needs to decide. Percent-complete is easy. It's also close to meaningless on a wall job — "72% complete" says nothing about whether drywall can start Monday on the third floor.
Before you trust any dashboard, figure out the three or four questions your morning actually turns on:
- What is scheduled to start in the next 5–10 working days, and is it actually ready to start?
- What constraints are open — material, RFI, inspection, predecessor work — and who owns each one?
- Which crews are committed where this week, and where are two trades about to collide in the same room?
- What did we commit to last week, and did we hit it?
If the dashboard answers those, the pretty gauges are optional. If it doesn't, the pretty gauges are noise. A short-interval schedule lives or dies on readiness and commitment reliability, so those are the numbers that earn the top-left corner of the screen — the spot your eye lands first.
Metrics That Actually Move a Job
Vanity metrics feel productive and change nothing. Here are the ones worth watching, and why.
Percent Plan Complete (PPC)
PPC is the single most honest number in look-ahead scheduling: of the tasks you committed to this week, what fraction did you finish as promised? Not "made progress on" — finished. A crew that's hitting 45% PPC is telling you the weekly plan is fiction, and the reasons why (call them variances) are the gold. Track PPC as a trend, not a snapshot. A single week at 60% is a bad week; eight weeks trending down from 80% to 55% is a job losing its grip on planning, and it usually shows up in PPC weeks before it shows up in the CPM.
Constraint aging
Every constraint has a clock. An RFI that's been open 18 days on work scheduled to start in 6 is a fire, not a line item. The most useful thing a dashboard can do is sort open constraints by days until the dependent work is supposed to start, not by date opened. That flips the list so the thing about to hurt you floats to the top on its own.
Look-ahead readiness
For everything in your two- or three-week window, what percentage is constraint-free and truly ready to release to a crew? If only 40% of next week's planned work is actually ready, you don't have a schedule, you have a wish list — and it's better to know that on Wednesday than to find out when the crew shows up Monday with nothing to do.
Trade collisions
The one metric nobody builds and everybody needs: how many locations have two or more trades planned in them at the same time this week? On a location-based plan this is easy to surface, and it's the difference between a smooth week and the electrician and the drywaller arguing over the same room at 7 a.m.
Visual Hierarchy: Design for a Six-Second Glance
A super doesn't study a dashboard. They glance at it between a phone call and a walk. If the most important thing on the screen isn't obvious in about six seconds, the design failed.
Real hierarchy means ruthless subordination. One or two things get to be big and bold — usually this week's PPC and the count of hot constraints. Everything else gets smaller and quieter. The mistake I see constantly is a grid of twelve equally-sized tiles, every one demanding attention, which means none of them get it. If everything's a headline, there's no headline.
Put the "act now" content top-left, where the eye lands first in any left-to-right layout. Push reference data — historical trends, cost curves, closed items — down and to the right where you go looking for it deliberately. And leave whitespace. A cramped dashboard reads slower than a clean one with fewer tiles, every time.
Color: A Signal, Not Decoration
Color on a jobsite dashboard should mean exactly one thing, consistently, everywhere: status. Red is a problem that needs a person today. Yellow is at risk. Green is on track. That's it. The moment you use red for "critical path" in one widget and "over budget" in another and "high priority" in a third, red stops meaning anything and people tune it out.
Two hard rules from painful experience. First, roughly one in twelve men on your crew is red-green colorblind, so never encode status by color alone — pair it with an icon, a label, or position. A red dot that's also flagged with a triangle and the word "OVERDUE" survives both colorblindness and a sun-washed tablet screen at 2 p.m. Second, use color sparingly. If 60% of the screen is colored blocks, the eye can't find the one red cell that matters. Color should be the exception that draws attention, not the wallpaper.
Real-Time Is Overrated; Trustworthy Is Everything
Everybody asks for "real-time." What they actually need is current enough to trust and honest about how fresh it is. A schedule dashboard that reflects this morning's plan is plenty real-time. What kills a dashboard is silent staleness — a number that looks live but is actually three days old because a sync failed and nobody was told.
The single most valuable, most ignored piece of dashboard UI is a small honest timestamp: "updated 7:42 a.m." When the field trusts that the plan on the wall matches the plan in the app, they use it. The first time someone gets burned by acting on a stale number, they quietly go back to the whiteboard and the group text, and you've lost them. Freshness you can see beats freshness you have to assume.
Drill-Down: Every Number Should Answer "Says Who?"
A summary number that can't be opened is an assertion, not information. When PPC reads 58%, the very next question is which tasks missed, and why — and if the dashboard can't get you from that number to the list of the seven activities that slipped and the variance reason on each, the number is just a topic for an argument. Good drill-down means one click from the headline to the underlying rows: this crew, this location, these dates, this constraint. If you have to leave the dashboard and dig through a separate report to answer the obvious follow-up, the summary saved you nothing.
Role-Based Views: Show People Their Own Work
The PM cares about cost trends and the four-week look-ahead. The foreman cares about what his crew does tomorrow and what's blocking it. The sub cares only about the locations and dates that involve his people. Feed all three the same screen and you've built something too crowded for the foreman and too shallow for the PM.
This is where a mobile companion earns its keep. A crew leader doesn't want a 12-tile executive dashboard on a phone in the rain — he wants his week, his locations, his handoffs, and the one thing that's blocked, big enough to read with gloves on. LookAheadWall's crew-leader app leans this way on purpose: the field view is the plan filtered down to the work that's actually theirs, not the whole enchilada shrunk to fit. Same underlying schedule, different altitude for different jobs.
Trends Over Snapshots
Any single number is a photograph; the trend is the movie, and the movie is where the story is. PPC at 65% this week is neutral. PPC that's been 82, 78, 71, 65 over four weeks is a job sliding, and you can intervene while it's still cheap. Sparklines next to the headline numbers — small inline trend lines — carry more decision-making value per pixel than almost anything else on the screen, because they turn "how are we doing" into "which way are we headed."
Don't Confuse Alerts With Notifications
Alert fatigue is real and it's fatal. A dashboard that pings you on every task update trains you to ignore it inside a week. Reserve true alerts for the short list that genuinely needs a human today: a constraint aging past the point where the dependent work can still start on time, a crew committed to a location that isn't ready, a collision between two trades in the same room. Everything else is a notification you can browse when you choose to. Fewer, sharper alerts get acted on. A firehose gets muted.
The Real Test
Here's the honest way to judge any construction dashboard, whatever the sales demo showed you. Watch what people do with it after 30 days. If the super still glances at it every morning before setting the crews loose, it's a real tool. If it's become the TV that plays in the corner while everyone runs the job off a whiteboard and a text thread, it's decoration — and no amount of gauges, color, or "real-time analytics" will fix that.
The dashboards that survive share one trait: they're built around the handful of decisions a superintendent actually makes each week — where do the crews go, what's ready, what's blocked, did we hit our commitments — and they answer those questions in a glance. Everything else is a nice-to-have. Get the four questions right and the screen becomes part of how the job runs. Get them wrong and you've bought a very expensive slideshow.