Construction quality control QA/QC program field guide

What a construction quality program is

A construction quality program is the documented system a contractor uses to build work that meets the contract and to prove it did. It has two halves that people constantly blur. Quality assurance is the planning and process that keeps defects from happening. Quality control is the inspection and testing that catches the defects that slip through anyway. You need both, and a program with only one of them fails in a predictable way.

The reason any of this exists is sequence. Construction buries its own work. The rebar disappears under the concrete, the concrete disappears under the slab finish, the waterproofing disappears behind the backfill, and the conduit disappears inside the wall. The cheapest defect on a project is the one caught before the next trade covers it. The most expensive is the one found by the owner after closeout, because by then the fix means demolition, schedule, and an argument about who pays.

Two sibling guides on this site show the principle at trade scale. The pre-pour rebar inspection guide is a hold point in action, because once the pour covers the steel you cannot see the cover or the laps again. The concrete strength testing guide is the testing leg, proving the placed concrete reached its specified strength. This guide is the program those checks live inside.

What is the difference between QA and QC?

Quality assurance is the system that prevents defects. Quality control is the inspection that detects them. QA is proactive and process focused: the quality plan, the procedures, the training, the approved submittals, the inspection and test plan written before work starts. QC is the hands-on activity in the field: the inspections, the checklists, the material checks, the lab tests, the daily report. One sets the work up to come out right. The other verifies that it did.

A clean way to keep them straight: QA is the plan and the process, QC is the inspection and the result. If you are writing a procedure or approving a product before anyone installs it, that is QA. If you are walking a wall with a checklist or pulling a sample for the lab, that is QC. Many people treat QC as a subset of QA, and that holds up, because the control activities only mean something inside a system that defined what acceptable looks like first.

The lean that matters is this. The more real QA you do, the less QC you need to catch, because the work shows up right. A program heavy on inspection and light on planning spends its life finding the same defects over and over and never asking why they keep happening.

Why a quality program matters: rework is pure loss

Rework is the most expensive thing a project does, and it produces nothing. You already paid labor and material to build the defect. Now you pay again to tear it out, pay a third time to rebuild it, and you lose the schedule the rework eats. None of that money shows up as a line item, which is exactly why it gets ignored until it is too big to ignore.

The buried defect is the trap. A wall that fails inspection while it is still open costs an hour and a conversation. The same defect found after the drywall, the paint, and the casework cost a fraction of what it costs to expose it again. The defect did not get worse. The work in front of it got expensive.

Then there is the callback and the dispute. A defect that reaches the owner becomes a warranty claim, a punch item that will not close, or a lawyer's letter. Most quality disputes are not really about the defect. They are about the absence of a record showing the work was ever checked. A program that prevents the defect and documents the check kills both problems at once. The prevention is cheap. The failure is not.

The cost of quality: prevention against failure

Cost of quality is the accounting that makes the case for a program, and it splits into four buckets. Prevention is what you spend to keep defects from happening: planning, training, the quality plan, getting submittals approved before purchase. Appraisal is what you spend to check: inspections, testing, lab fees, surveys. Those two are the cost of conformance, the money you spend on purpose. Internal failure is rework caught before handover. External failure is the callback, the warranty repair, the damage claim after the owner has the building.

The lesson from the numbers is consistent across industries. Failure cost dwarfs prevention cost. Money moved into prevention and appraisal comes back as a larger reduction in failure, and the total cost of quality drops. Spending nothing on prevention does not save money. It moves the spend into the failure column and multiplies it.

The iceberg is the useful picture. The visible failure cost is the rework you can see and price. Under the waterline sit the hidden costs: the schedule slip, the trades stacked behind the fix, the inspector who now witnesses everything because trust is gone, the reputation that loses you the next job. On construction work, that hidden base routinely runs several times the visible tip. The exact ratio varies with the project and how you count it, so treat the principle as solid and the precise figure as project specific.

The project quality plan

The quality plan is the written document that says who is responsible for quality, what gets checked, and how. On federal and many institutional jobs it is called the Contractor Quality Control plan, the CQC plan, and it is a contract deliverable you cannot start work without. On private work it may be lighter, but the good ones cover the same ground.

A working plan names the quality organization and the people in it, with the QC manager's authority spelled out, including the authority to stop work. It lists the definable features of work, the activities the program will control. It points to the standards each feature is built to. It contains or references the inspection and test plan. It sets the procedures for submittals, material control, inspections, testing, nonconformances, and documentation. And it says how records get kept and turned over.

The plan is only worth the paper if the field runs it. A binder that gets approved and then sits in the trailer while the work is built on habit is not a quality program. It is a compliance prop. Write the plan you will actually follow, scaled to the job, and treat the contract requirements as the floor, not the target.

Building to the spec, the contract, and the submittals

Quality means conformance to requirements, so the first job of any program is to pin down what the requirements are. Acceptable is not an opinion. It is defined by the contract documents: the drawings, the project specifications, the codes the jurisdiction adopted, and the approved submittals. When a crew and an inspector disagree about whether work is good, the answer is in those documents, not in either person's experience.

The order of precedence matters, and the contract usually states it. The project specification can be tighter than the code minimum, and where it is, the spec governs. An approved submittal becomes the standard for that product or assembly, which is why the approved shop drawing, not the salesman's catalog, is what you inspect against. A reference standard like an ACI or ASTM document carries only the parts the spec invokes.

Keep document control tight, because building to a superseded drawing is one of the most expensive errors there is. The crew has to be working from the current revision and the approved submittal, not the bid set or a marked-up print from three weeks ago. If you cannot tell at a glance which revision is in someone's hands, that is the defect waiting to happen.

What is an inspection and test plan (ITP)?

An inspection and test plan is the matrix that says, for each activity, what gets inspected or tested, against which standard, how often, who does it, what record proves it, and where the inspection gates sit. It is the working heart of quality control, the document the field actually runs off. A quality plan describes the system. The ITP turns it into a row-by-row schedule of checks tied to the work.

Each row of an ITP carries a consistent set of columns. The activity or step. The reference: the spec section, drawing, or code clause that defines acceptance. The acceptance criteria themselves. The method, whether it is a visual check, a measurement, or a lab test. The frequency or sampling rate. The responsibility, naming who inspects and who else must be notified. The control point, marked as a hold or witness point. And the record that closes the row out.

Build the ITP from the specification, feature by feature, so nothing required gets dropped and nothing gets buried before its check. The sibling guides are two rows made real: a pre-pour rebar inspection is a hold point on the reinforcing activity, and cylinder strength testing per ASTM C39 is the test on the concrete activity, at a sampling frequency the spec sets. Scale the detail to the risk; a critical structural pour earns more rows than a sidewalk.

What is a hold point versus a witness point?

A hold point is a stop. Work cannot proceed past it until the required party has inspected and released it in writing. A witness point is a notification. You tell the inspecting party the work is happening so they can attend if they choose, but if they decline or do not show within the agreed window, you may proceed. Both are inspection gates set in the ITP. The difference is whether the work waits.

Hold points go where the work is about to become invisible or irreversible. The classic is the pre-pour inspection: once concrete covers the rebar, the cover, the laps, and the embeds are gone for good, so the reinforcing has to be released before the pour. Waterproofing before backfill, structural connections before fireproofing, and underground utilities before cover are the same logic. Bury work past an unreleased hold point and you have created an NCR and probably an excavation.

Set the notification window for witness points in the plan so it is not an argument on the day, commonly a day or two ahead depending on the contract. The discipline that makes hold points work is simple and unforgiving. The crew has to know the gate is there before they reach it. A hold point nobody briefed is a hold point that gets blown through, and the cost of that lands after the next trade is already on top of it.

First-work and the mockup

The first-work check, sometimes a formal mockup, is where you inspect and approve the first unit of a repeated operation before the crew mass-produces it. Approve one good masonry panel, one good window installation, one good rebar mat, and it becomes the benchmark every following one is measured against. Catch the defect on unit one and you fixed one unit. Catch it on unit fifty and you own fifty.

A mockup does double duty. It sets the visual and workmanship standard the owner and architect sign off on, so aesthetic disputes get settled on a sample panel instead of on the finished building. And it proves the means and methods work before they are committed at scale: the right mortar joint, the correct flashing sequence, a finish that actually meets the sample. On systems with a weather or warranty consequence, the mockup is also where the installer confirms the detail in the field, because the detail on the drawing usually assumes conditions the jobsite does not give you.

The mistake is skipping it to save a day. The crew runs the whole elevation, the architect rejects the joint tooling, and now the rework is the elevation, not a panel. First-work is the cheapest place to be wrong.

Preparatory, initial, and follow-up control

The three-phase control method comes from the US Army Corps of Engineers and shows up on most federal and many institutional jobs, but the structure is good practice anywhere. It applies three checks to each definable feature of work, spaced across its life so problems get caught before they multiply.

The preparatory phase happens before work starts on the feature. It is a meeting that confirms the submittals are approved, the materials on hand match them, the crew understands the spec and the acceptance criteria, the preceding work is ready, and the inspection and testing for the feature is set. The initial phase happens as the first work goes in. It is the hands-on check of the first piece, the same idea as first-work, confirming the workmanship standard is set before the crew runs. The follow-up phase is the ongoing monitoring that the standard holds as the feature is produced.

The value is in the spacing. Preparatory catches the planning gap, missing submittal, wrong material, unready substrate. Initial catches the workmanship problem on day one. Follow-up catches drift. Run all three and most defects surface while they are still a single unit and a conversation, which is exactly where the cost of quality wants them. Document each phase, because on a CQC job the daily report has to show they happened.

Material verification against the approved submittal

Material verification confirms that what showed up on the truck is what was approved, before it goes into the work. The submittal process exists so the engineer and architect approve a product on paper first. The receiving check closes the loop by proving the delivery matches that approval. Skip it and an approved submittal becomes meaningless, because the wrong product gets installed under cover of a paperwork trail that says everything is fine.

What you check depends on the material, but the pattern holds. Compare the product, model, grade, and size against the approved submittal. Pull the mill certificate or certificate of compliance for steel, rebar, and structural products and confirm the heat or batch matches the documentation. Check the concrete delivery ticket against the approved mix design before the truck discharges. Look at the physical condition: damage, contamination, expired shelf life on sealants and adhesives, correct storage. Reject and tag anything that does not conform so it cannot drift into the work by accident.

The quiet failure here is the substitution nobody approved. A supplier ships an equivalent that was never submitted, the crew installs it because it looks right, and the deviation surfaces during closeout when the certs do not line up. Catch it at the gate, not in the turnover binder.

Running the QC inspections

The inspection is where quality control meets the work. A QC inspection compares the installed work against the acceptance criteria using a checklist built from the spec, so the inspector checks what the contract requires rather than what catches the eye. The checklist is what keeps inspections consistent between people and between days, and it is what makes the inspection defensible later.

Inspect against the standard, in the right sequence, and at the frequency the ITP sets. For covered work, the inspection is the hold point that releases the next operation, which is why timing matters as much as content. An inspection done after the cover is too late to be an inspection. It is an excavation report.

The daily QC report ties it together. At minimum it records the work inspected, the results, any deficiencies and their status, materials received, tests taken, the weather, and which phase of control happened on which feature. On a CQC job the report is a contract record and the owner's representative reads it. Even where it is not required, the daily report is the contemporaneous record that answers, months later, whether a given piece of work was ever actually checked. Write it the day of, not from memory the week after.

Testing: the lab, the field, and the independent eye

Testing is the quantitative leg of quality control, the part that produces a number instead of a judgment. Some properties cannot be seen, only measured, so the program leans on tests to confirm them: concrete compressive strength, soil compaction, weld integrity, bolt tension, sprinkler pressure, duct leakage. The spec sets which tests apply, the acceptance criteria, and the sampling frequency, and the ITP carries them as rows.

Most spec-required testing is done by an accredited independent laboratory, not the installing contractor, so the result carries weight. The concrete strength testing guide on this site walks the full cycle for one of these: sampling fresh concrete, casting and curing cylinders per ASTM C31, breaking them per ASTM C39, and judging the result against the ACI 318 acceptance rule. Soils, asphalt, masonry, structural steel, and firestopping each have their own test methods and their own accredited testers.

Two field habits matter. Sample at the point that represents the work, because a sample taken in the wrong place tests nothing useful. And track the result against the acceptance criterion the moment it lands, so a failing test triggers the nonconformance process while the work is fresh and the truck or the crew is still reachable, not weeks later when the report finally gets read.

What is a nonconformance report (NCR)?

A nonconformance report is the formal record that documents work, material, or testing that does not meet the contract requirement, and it tracks that defect to a closed resolution. It is raised when something fails an inspection or a test, when the wrong material gets installed, or when work was buried past a hold point without release. The NCR is how a quality program proves it found its own problems and dealt with them instead of hiding them.

The structure is consistent. The NCR identifies the nonconformance specifically, with location, the requirement it violates, and supporting evidence including photos. It segregates and contains the affected work so it does not spread or get built over. It carries the defect through a disposition decision. And it captures the corrective action that keeps it from recurring, then verifies the fix before it closes.

The blunt rule is that you write the NCR. Hiding a nonconformance is the single worst thing a quality program can do, because the defect does not disappear, it just goes into the building unrecorded, waiting to surface during warranty when it is most expensive and least defensible. A documented NCR with a sound disposition is a sign the system works. A clean file on a messy job is a sign the system is being bypassed.

The NCR disposition

Disposition is the decision about what to do with nonconforming work, and there are four standard answers. Rework brings the work back into full conformance with the original requirement. Repair brings it to a usable condition that does not fully meet the original spec, and it requires engineering approval because someone is accepting less than what was drawn. Use-as-is accepts the nonconformance with no physical change, on the judgment that the deviation does not affect form, fit, function, or strength. Reject removes it.

Who decides depends on the disposition. Rework and reject usually sit with the contractor. Repair and use-as-is on anything structural or contractually significant require the engineer of record's written acceptance, and often the owner's, because they are agreeing to take the building with a known deviation. That approval is the record that protects everyone later, so get it in writing before the work moves on, not as a verbal in the field that nobody can find at closeout.

Treat use-as-is with discipline. It is the right call when the deviation genuinely does not matter, and it is a quiet way to launder defects when it is used to avoid rework on things that do. The engineer's stamp is what separates the two, and the project specification controls when it is required.

Corrective action and the root cause

Fixing the defect is not the same as fixing the cause. Disposition handles the unit in front of you. Corrective action handles the reason it happened, so the next one does not. A program that only ever disposes of defects, and never asks why they keep appearing, will see the same nonconformance on a steady cycle and call it bad luck.

Root cause is the discipline that separates the two. When the same defect shows up more than once, push past the symptom to the cause: a wrong detail that keeps getting built, a crew that was never briefed on the spec, a supplier that keeps shipping off-grade material, a hold point that is not being honored. Simple tools like asking why repeatedly until the real cause surfaces, or a cause-and-effect breakdown, are usually enough on construction work. The corrective action targets that cause, and a preventive action checks whether the same gap exists somewhere else on the job.

Trend the NCRs to find the cause worth chasing. If a single trade or a single detail generates a third of your nonconformances, that is where root-cause work pays. The productivity and quality of the whole job improves more from killing one recurring defect than from disposing of fifty one-off ones.

Documentation and the proof

The quality program produces a defect-free building and a record that proves it, and the record is half the deliverable. The daily QC reports, the inspection checklists, the test results, the material certs, the NCRs and their dispositions, and the photos of covered work are what answer the question that always comes later: was this checked, and by whom. Quality work with no record is, in a dispute, indistinguishable from no quality work at all.

Photos earn their place in the modern file. The pre-pour photo of the rebar, the shot of the waterproofing before backfill, the embed before the concrete, these are the evidence that the hold point was real and the work behind the finish is right. They turn an inspection from a claim into proof.

This is where a field tool earns its keep. FieldOS captures the inspection, the checklist, the photo, the test result, and the NCR at the point of work, time-stamped and tied to the location, and assembles the turnover package as the job runs instead of in a closeout scramble. The record built the day the work happened is worth far more than the one reconstructed from memory at substantial completion, both as proof and as the thing that makes the daily reporting survivable.

Owner QA over contractor QC

There are usually two layers of quality on a project, and they are not the same job. The contractor runs quality control: building the work and inspecting its own production. The owner runs quality assurance over that: oversight, audits, and independent verification that the contractor's QC is real and working. On federal jobs this is the explicit split, contractor QC and government QA, and the same structure appears on most large private and institutional work.

The owner's QA can take several forms: a resident engineer or owner's representative reviewing the daily reports, periodic audits of the QC program, witness of hold points, and independent or third-party inspection and testing that runs parallel to the contractor's. The point of the second layer is that self-inspection has a built-in conflict, and an outside eye keeps the program honest.

For the contractor, the practical effect is that the QC program has to stand up to outside review. A program that exists to satisfy the owner's QA and nothing more is thin, but one that produces records clean enough to survive an audit is usually a program that is actually controlling the work. Hedge to the contract: the exact division of QA and QC roles is defined there and varies by delivery method.

Code special inspections (IBC Chapter 17)

Special inspections are a separate, code-mandated layer that sits on top of the contractor's QC and the building department's own inspections. Under the International Building Code, Chapter 17 requires that certain work, mostly structural, be inspected by a qualified special inspector working for an approved agency, independent of the contractor installing it. This is not the same as the city inspector's framing or final, and it is not the contractor's QC. It is a third thing the code requires for higher-risk elements.

The list of what needs special inspection is set by Chapter 17 and the project's statement of special inspections, which the design professional prepares and the building official approves as a condition of the permit. Common items include structural concrete and reinforcing placement, structural steel and its welds and high-strength bolting per AISC 360, masonry, soils and deep foundations, and sprayed fire-resistant materials. The special inspector reports nonconforming work to the building official, the contractor, and the design professional, which is why an uncorrected item can hold up the certificate of occupancy.

Two practical notes. The required inspections and inspector qualifications are set by the adopted code edition and local amendments, so confirm them for the jurisdiction rather than assuming. And the registered design professional in responsible charge can sometimes serve as the approved agency for their own design, where they qualify, but that is governed by the code, not by preference.

Holding subcontractors to the quality plan

Most of the work on a building is built by subcontractors, so a quality program that stops at the general contractor's own forces controls almost nothing. The plan has to flow down. The subcontract should require each sub to work to the project quality plan, to support the inspection and test plan and its hold points, and to provide the submittals, certs, and records the program needs.

In practice this means the subs attend the preparatory meetings for their features, their work hits the same hold points, and their material gets the same receiving check. Where a sub runs its own QC, as specialty trades like structural steel, fire protection, and curtain wall often do, the program coordinates with theirs rather than ignoring it, and the records still land in the project file.

The coordination failure shows up at the interfaces. One trade's work becomes the next trade's substrate, and the defect that the first sub left gets built over by the second, who had no reason to inspect it. The quality plan's job at those handoffs is to put a check on the boundary, so the receiving trade is not unknowingly building on a nonconformance. Define those interface inspections in the ITP, because nobody owns them by default.

How quality reduces the punch list

The punch list at the end of a job is, in large part, a measure of how well quality control ran during it. Every item on it is a defect that was built and not caught until the owner's walk. A program that inspects work as it goes, catches defects at first-work, and holds the gates produces a short punch list, because the defects got fixed when they were cheap instead of accumulating for the final walk.

The expensive punch items are the ones that should have been hold points. A finish defect is a touch-up. A waterproofing miss found at substantial completion is a tear-out. The closeout guide territory, the systems that have to be commissioned and the documents that have to be turned over, runs smoother when the quality records were built along the way rather than assembled in the last two weeks.

There is a schedule argument here too. A long punch list does not just cost the rework. It holds the retainage, delays the certificate of occupancy, and keeps a crew on a job that should be finished. Quality during the work is the cheapest way to get off the job on time.

Quality is built by the crew, not inspected in

You cannot inspect quality into a building. Inspection finds defects after they exist, and a program that relies only on catching them is always one inspection behind the work. The work comes out right because the crew that placed it built it right, understood the standard, and cared about the result. The quality plan, the ITP, and the inspections support that. They do not replace it.

The practical version of this is that the program's real job is to put the crew in a position to do it right the first time. The preparatory meeting that confirms the crew knows the spec, the approved submittal in their hands instead of a guess, the first-work check that sets the standard before they run, the material that was verified before they touched it, all of that is the program building quality at the source. Inspection then confirms what good work already produced.

A foreman who treats the inspector as the enemy runs a program that finds defects. A foreman who treats the standard as their own runs one that prevents them. The second job is faster, cheaper, and the one people want to be on. Culture is what turns a binder into a building.

The metrics that drive improvement

A quality program that does not measure itself cannot improve, and the metrics worth tracking are few. Count NCRs by trade, by feature, and by root cause, so the recurring defect surfaces instead of hiding in the noise. Track rework cost, even roughly, because the dollar figure is what makes the cost-of-quality argument to people who control the budget. Track first-time quality, the share of work that passes inspection without rework, as the headline number that says whether the program is working.

The point of the metrics is the trend, not the snapshot. A first-time-quality rate that climbs as the job runs says the preparatory meetings and first-work checks are landing. An NCR count concentrated in one trade says where to send the next root-cause effort. The numbers turn the program from a set of checks into a feedback loop, which is the process-and-improvement idea at the center of ISO 9001 brought down to a single project.

A field tool makes the metrics free. When FieldOS captures every inspection, test, and NCR as it happens, the trend reports build themselves instead of requiring someone to tally a binder. Measure what the program produces, feed it back into the next preparatory meeting, and the defects that keep happening stop happening.

What the quality records have to include

The quality record is the program's memory, and it has to be complete enough that someone who was never on the job can reconstruct whether the work met the contract. That means the quality plan and its revisions, the inspection and test plan, the inspection and test records tied to the features they cover, the material certifications and approved submittals, the NCRs with their dispositions and corrective actions closed out, and the photographic evidence of covered work.

Keep the records contemporaneous and traceable. A test result has to tie to the placement it represents, an NCR to the location and the requirement it violated, a hold-point release to the work it let proceed. Records assembled after the fact, with gaps papered over, are worse than useless in a dispute because they read as exactly what they are.

This whole file is the turnover deliverable as much as the building is. Capturing it in a field tool like FieldOS as the work happens, rather than reconstructing it at closeout, is the difference between a turnover package that proves the job and a scramble that papers over the holes.

Where quality programs break down

Most failed programs fail in the same handful of ways, and they are recognizable from across the trailer. No real quality plan, so quality runs on the personality of whoever happens to be on site. No ITP or no hold points, so the inspections are not tied to the work and defects get buried before anyone checks them. No first-work or mockup, so the crew mass-produces a defect that one early check would have caught.

The two that do the most damage are quieter. Hiding NCRs instead of writing them, which puts unrecorded defects into the building and destroys the program's credibility the moment one surfaces. And fixing the defect without fixing the cause, so the same nonconformance comes back on a cycle while everyone treats each instance as new. A program can have a thick binder and still do both of these. The binder is not the program. What the field does is.

Common mistakes

• Running the job with no real quality plan, so quality depends on who is on site that day.
• No inspection and test plan or no hold points, so defects get buried before anyone inspects them.
• Skipping the first-work or mockup check and mass-producing a defect one early check would have caught.
• Hiding nonconformances instead of documenting them, which puts unrecorded defects into the building.
• Not verifying delivered material against the approved submittal and mill certs at the receiving gate.
• Fixing the defect but never chasing the root cause, so the same nonconformance keeps coming back.
• Building to a superseded drawing because document control was loose.

Field checklist

Standards and references

The contract is the first authority. The project quality requirements, often a specification section dedicated to quality control or a required CQC plan, define what the program has to do and what it has to deliver, and they govern over any general guidance. Build to the drawings, the project specifications, the adopted codes, and the approved submittals, in the order of precedence the contract sets.

Beyond the contract, several frameworks shape construction quality. ISO 9001 gives the quality-management principles the whole field rests on, including the process approach and the plan-do-check-act improvement cycle. The inspection-and-test-plan and hold-point method, and the three-phase preparatory, initial, and follow-up control, come from the US Army Corps of Engineers and NAVFAC construction quality management and are widely adopted beyond federal work. The International Building Code, Chapter 17, governs code-required special inspections and the statement of special inspections, with reference standards like AISC 360 for structural steel and the ACI documents for concrete invoked as the spec calls them.

Cite the standard that actually controls the point, and hedge the rest to the project. The required special inspections, the code edition, and the inspector qualifications are set by the adopted code and local amendments. The acceptance criteria, the testing frequency, and the disposition authority are set by the project specification and the engineer of record. The principles here are stable: quality assurance prevents and quality control detects, build to the spec, the ITP with its hold points and first-work checks catches the defect early, and every nonconformance gets documented and fixed at the root.

Terms and definitions

The vocabulary of construction quality is precise, and using the words loosely is how QA and QC get confused in the first place. The terms below are the ones a quality program runs on.

Quality assurance (QA)

The planning and process that prevents defects: the quality plan, procedures, training, and approved submittals set before work starts

Quality control (QC)

The inspection and testing that detects defects in the work as it is built

Inspection and test plan (ITP)

The matrix listing each activity, its standard, acceptance criteria, method, frequency, responsibility, hold or witness point, and record

Hold point

A mandatory gate where work stops until the required party inspects and releases it in writing

Witness point

A gate where the inspecting party is notified and may attend, but work may proceed if they do not within the agreed window

First-work / mockup

Inspection and approval of the first unit of a repeated operation, which becomes the benchmark for the rest

Nonconformance report (NCR)

The formal record of work or material that does not meet the requirement, tracked through disposition and corrective action to closure

Disposition

The decision on nonconforming work: rework, repair, use-as-is, or reject, with engineer approval where required

Cost of quality

Prevention plus appraisal (conformance) against internal plus external failure (rework and callbacks); failure cost dwarfs prevention

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• How is concrete strength tested?
• What is the minimum concrete cover for rebar?
• What is an inspection and test plan?
• What is construction dewatering?
• What is construction layout?