Five Construction Best Practices to Insist on
What we've learned across 40+ projects and $1+ billion in development
These are the basics we always require.
Most construction “best practices” lists fill pages with obvious advice about communication and planning. Everyone nods along and ignores them because they sound like common sense.
The practices that actually keep projects on track are more specific than that. They’re the unglamorous, daily disciplines that experienced contractors execute automatically and that struggling projects skip because they look like administrative overhead.
From working on dozens of projects, here are five construction practices every one of our projects must follow.
1. Keep a clean jobsite.
Cleanliness on a jobsite has nothing to do with appearance. Poor material staging sends workers hunting for things that should take two minutes to find. Debris gets in the way and slows trades down. Overflowing dumpsters leave trades with nowhere to dispose of waste. Every one of those problems shows up in the schedule.
What “clean” actually means:
Daily cleanup at end of shift (not weekly, daily)
Designated material staging areas that don’t block access
Dumpster schedule that prevents overflow
Clear pathways for movement and egress
Proper storage of hazardous materials
Cleanup requirements belong in the construction contract, enforced through weekly site inspections. Projects that maintain clean sites consistently finish faster because trades aren’t getting in each other’s way or losing track of stored materials that often need to be reordered.
We’ve seen projects lose entire days because a delivery truck couldn’t access the site due to poor staging. We’ve seen injuries from trips and falls that basic housekeeping would have prevented. The time lost from a messy site always costs more than the time spent keeping it clean.
This is also a signal of a GC/CM who is on top of the details because they understand the upstream effects.
2. Require sign-in sheets and actually use them.
Every person on site should sign in every day. The CM tracks who’s there, what they’re working on, and whether the project has the manpower to hit the schedule.
What gets tracked:
Company name
Number of workers
Time in/time out
Three reasons this matters:
First, liability. If someone gets injured on site, the record of who was there and who they work for is essential. Insurance will require this documentation.
Second, labor loading. If the schedule shows 15 framers on site this week but the sign-in sheet shows 6, there’s a problem. The sign-in sheet is the early warning system for schedule slippage.
Third, quality control. If an issue comes up with specific work, the record shows which trade did it and when.
The CM should review these sheets weekly as part of schedule update meetings. Are the labor counts matching projections? Are trades showing up when the schedule says they should? This simple document catches problems before they become delays.
3. Control the change order process.
Budgets die here, not from one big surprise but from dozens of small change orders that compound because no one was paying attention.
What counts as a legitimate change:
A legitimate change order addresses work that wasn’t in the original scope, like unforeseen conditions, owner-requested modifications, or code requirements that emerged during permitting. These are real changes that deserve compensation.
What doesn’t count:
Contractor errors, poor coordination between trades, work that was clearly in the drawings but missed in the bid. These are mistakes, not changes, and the owner shouldn’t pay for them. This is what the contingency is for.
The approval process:
Every change order request needs to answer two questions before approval.
Is this actually a change order?
The first step is confirming whether the work was part of the initial contract. If it was, the next check is whether it was raised within the contractually required timeframe. A change order appearing months after the work is complete is inexcusable. This is why we always include CO notification language in the contract.
If something is presented as a change order but shouldn’t be, the CM needs to ask questions. Common explanations:
The GC bought the scope, but the subcontractor forgot
The GC bought the scope from a different trade than the one submitting the CO
The GC forgot to buy the scope
If the GC forgot the scope, that comes out of the GC’s contingency. If a trade forgot to buy scope that was in their contract, the GC resolves that with the subcontractor, but it doesn’t become a change order to the owner, and if the scope is covered elsewhere, then it’s on the GC to detangle who is responsible and ensure the work is completed as bought.
Is the cost accurate?
This isn’t about “nickel and diming,” it’s about ensuring the cost breakdown holds up. Most inaccuracies are honest mistakes that get corrected once someone identifies them.
Common mistakes we often see in change order cost breakdowns:
Material from a different project charged to this one
Double counting
Incorrect material takeoffs versus field conditions
Labor rates that don’t match the contract
Estimated labor charges instead of actual
Unnecessary trades added to the change order
Wrong markup versus what the contract specifies (the most common)
The CM reviews the breakdown, verifies the change is legitimate, checks the math, and approves or rejects it within a defined timeframe. Without this process, the owner pays for work already in the contract, or pays incorrect amounts because no one checked.
We’ve seen projects go 20% over budget purely from change orders that either shouldn’t have been approved or carried incorrect cost breakdowns.
The key is keeping change order review current. Unwinding months of incorrect change orders is time-consuming, and it carries real project risk. Backcharging contractors is difficult under any circumstances, and for smaller trades without the balance sheet to absorb a chargeback from months prior, it can trigger a default.
4. Structured weekly meetings for schedule and log review.
Monthly schedule updates aren’t enough. By the time a schedule problem surfaces in a monthly meeting, weeks have already been lost. Email threads on submittals and RFIs compound this because tracking who has responded, and when, becomes nearly impossible at volume. Software exists to solve exactly this problem, but the logs only work if someone is actually reviewing them. Weekly meetings solve both problems.
What we cover:
Two-week lookahead. The master schedule tells you where the project is supposed to be. The lookahead tells you whether it’s actually getting there. It shows specific tasks by trade, material delivery dates, required manpower, coordination requirements between trades, and scheduled inspections. This is the document the superintendent uses to coordinate day-to-day, and it’s the earliest place a schedule problem becomes visible.
Submittal, RFI, and change order logs. Every open item gets tracked with the date it was submitted, who holds responsibility for review, when a response is due, and current status. Logs should also reflect how long each item has been open and, where warranted, a priority designation. Priority should be used selectively and tied to real schedule impact rather than applied broadly, otherwise it loses meaning. In weekly meetings the team reviews open items approaching their response deadlines, items that have already exceeded them, and any patterns in where the approval process is bottlenecking.
Open items log. Design questions, owner decisions, and anything the GC or CM needs clarification on to move forward. Items stay on the list until they are confirmed resolved. This log is frequently overlooked, but unresolved open items are one of the most consistent sources of delay we see across projects.
What to watch for:
If the master schedule shows framing complete by Friday and the lookahead shows it stretching into next week, there is a discrepancy that needs an explanation. If drywall is scheduled before MEP rough-in is complete, someone is not paying attention to sequencing. If a submittal has been sitting with the architect for fifteen days against a ten-day contract requirement, that needs to be resolved before it becomes a schedule event.
Why this matters:
Projects with current lookaheads and active log review run better because the team is forced to think ahead about materials, manpower, and coordination rather than react to problems after they develop. The weekly cadence also eliminates the “we’re waiting on” excuse. When response times are documented and reviewed consistently, the record is clear. Either the response is late or it isn’t. Either the decision was made or it wasn’t. We have seen projects slip by months because nobody was tracking that, and by the time the delay was visible, it was too late to recover cleanly.
5. Daily site reports (and actually read them).
Every day, the superintendent documents what happened on site, from weather conditions, manpower counts, work completed, to deliveries received, and problems encountered.
When a change order dispute comes up six months into construction, the daily reports can be valuable documentation, if they exist. “We had to stop work for three days due to unforeseen conditions” means nothing without reports showing when work stopped and why. When a schedule claim arises, the reports show what actually occurred versus what the schedule projected, proof of what caused delays.
(Aside: having said that, there is no excuse to delay issuing change orders for cost and/or time in a timely manner.)
What gets documented:
Weather conditions (critical for exterior work and schedule claims)
Manpower by trade
Work completed (specific tasks, not just “framing continues”)
Material deliveries
Issues, delays, or problems
Visitors to the site
The difference between useful reports and paperwork is specificity: “Framing continues” tells you nothing; “Completed second floor framing, started third floor deck” is more useful.
The CM reviews these reports at least weekly, if not daily. They show in real time whether progress matches the schedule and catch problems early. If reports show the same issue three days in a row, someone needs to intervene.
This is why we ask to be copied on daily reports across all active projects. On jobs that are running well, we spot check them. On jobs that aren’t, they become one of our primary diagnostic tools for understanding what’s happening on the ground and where intervention is needed.
If you do nothing else:
Insist on a clean site, meet weekly, and control the change order process. Those three practices prevent most of the problems we see on struggling projects. Sign-in sheets and daily reports matter most when something goes wrong and the project needs documentation to defend its position.
Good contractors execute all of these without being asked. That’s actually the point. How a contractor responds to these requirements early in a project tells you a great deal about how they’ll manage it when things get complicated.
If you are interested in investing in our fund, let us know via the investor interest form below.
AND
We are looking for long-term capital partners interested in a 5 to 10-year hold and looking for investment opportunities in the $10-$50mm range.
OR
If you have a development site that fits one the following criteria, we’d love to see it:
1 - In Manhattan with at least 40,000 buildable square feet zoned for residential use as of right, or a good candidate for an office-to-residential conversion
2 - Mixed-use, residential, or hospitality site in the mountain west
3- Destination or resort-anchored site with at least 50k of buildable ft
Most importantly, we want to meet and begin building relationships with people who are interested in what we are doing and who believe in how we approach the work, regardless of interest in our current deals.
Send us a note if you’d like to chat!
In addition to our development work in NY, MADDPROJECT provides fee development and professional project management services across the US.
Need help and don’t know where to start?
Have a specific subject you want to see us tackle in a future Newsletter?
Drop us a note!
