CAPA in Manufacturing: A Guide to Corrective and Preventive Action
Introduction
Your customer rejects a batch of parts. Dimensions are out of spec. You tell the operator to adjust the machine, rerun the parts, and ship replacements.
Two months later, the same customer rejects another batch. Same dimensions. Same machine. Different operator this time. You adjust and rerun again.
This cycle has a name. It's called fixing the symptom without fixing the cause. And it's expensive. Manufacturers typically spend around 20% of total sales revenue on the cost of poor quality, including scrap, rework, warranty claims, and the labor burned chasing the same problems over and over.
CAPA (corrective and preventive action) is the formal process for breaking that cycle. Instead of patching the immediate problem and moving on, CAPA forces you to find the root cause, fix it, and verify that the fix actually worked. Instead of waiting for the next failure, it asks what else might go wrong and addresses it before it does.
Most guides on CAPA are written for pharmaceutical companies or medical device manufacturers with dedicated quality departments and full-time compliance staff. This one is for small and mid-size manufacturers who need a working CAPA process without the bureaucratic overhead.
What is CAPA?
CAPA stands for corrective and preventive action. It's a structured approach to quality problems that goes beyond the immediate fix to address the underlying cause.
The two halves do different things:
Corrective action starts after something has already gone wrong. A part fails inspection. A customer returns a defective product. A process drifts out of control. Corrective action investigates why it happened, fixes the root cause, and puts controls in place so it doesn't happen again.
Preventive action starts before something goes wrong. You notice a trend in your inspection data. A risk assessment identifies a weak point in your process. An operator reports that a fixture is wearing unevenly. Preventive action addresses the risk before it produces a nonconformance.
The distinction matters. Corrective action is reactive: something broke, and you're making sure it stays fixed. Preventive action is proactive: nothing broke yet, but the data says it will.
Together, they form a feedback loop. Your corrective actions generate data about what goes wrong. That data feeds your preventive actions, which reduce the number of corrective actions you need in the future. A good CAPA system gets quieter over time, not louder.
Where CAPA fits in your quality system
CAPA is one part of a broader quality management system. It connects to your inspection processes, your nonconformance handling, your manufacturing traceability, and your management reviews.
If you track first-pass yield, CAPA is how you improve it. If you track scrap rates, CAPA is how you reduce them. If you track customer complaints, CAPA is how you prevent the same complaint from showing up twice.
It's not a standalone program. It works when it's connected to the data your shop already generates.
Why CAPA matters for manufacturers
Quality problems cost money. That's obvious. What's less obvious is how much of that cost comes from repeat problems rather than new ones.
The cost of fixing symptoms
When you fix a symptom, you incur the cost of the fix every time the problem recurs. The scrapped material, the rework labor, the expedited shipping, the inspection time, the customer communication. None of that is a one-time expense if the root cause is still there.
A machine shop that reworks the same tolerance issue four times a year at $3,000 per incident isn't dealing with four problems. It's dealing with one problem it never actually solved, at $12,000 a year. CAPA converts that recurring cost into a one-time investigation and a permanent fix.
Regulatory and customer requirements
If you work in aerospace, medical devices, automotive, or food manufacturing, CAPA isn't optional. ISO 9001 requires organizations to take action on nonconformities and evaluate whether corrective action is needed. AS9100 adds requirements specific to aerospace. ISO 13485 makes CAPA a formal, documented procedure for medical devices. The FDA considers "insufficient corrective and preventive action procedures" one of its most common compliance findings during inspections.
Even outside regulated industries, OEMs and prime contractors increasingly expect their supply chain to have a functioning CAPA process. If your customer's quality questionnaire asks about your corrective action procedures and you don't have a clear answer, that's a risk to the relationship.
Production data you can act on
CAPA generates useful data as a byproduct. When you investigate a root cause, you learn something about your process. When you track corrective actions over time, patterns emerge. Maybe 40% of your CAPAs trace back to the same work center. Maybe most nonconformances happen during changeovers. Maybe a specific supplier's material shows up in your quality records disproportionately.
This is the kind of information that connects directly to production scheduling decisions, OEE improvements, and job costing accuracy. The shop that runs CAPA well doesn't just have fewer quality problems. It has better visibility into why problems happen.
The CAPA process, step by step
The core process follows a logical sequence: identify the problem, find the cause, fix it, and verify the fix worked. Here's how each step works in a manufacturing setting.
1. Identify the problem
A CAPA starts with a trigger. Common triggers include customer complaints, failed inspections, audit findings, nonconforming material, equipment failures, and trend data that shows a process drifting.
Not every problem needs a full CAPA. A one-time mistake that an operator catches and corrects doesn't require a formal investigation. CAPA is for systemic issues, recurring problems, or significant events that could affect product quality or customer satisfaction.
The first step is documenting what happened in specific, measurable terms. "Parts were out of spec" isn't enough. "Lot 20260310-A, Job 4580, bore diameter measured 0.503" on 12 of 50 parts, tolerance is 0.500" +/- 0.001"" gives you something to investigate.
2. Contain the problem
Before investigating the root cause, contain the immediate impact. Quarantine suspect material. Segregate affected inventory. Notify customers if defective product may have shipped. This is the correction, not the corrective action. You're stopping the bleeding while you figure out what caused the wound.
Good traceability makes containment faster. If you can trace the affected lot to specific jobs and customers, you quarantine only what's affected instead of pulling everything off the shelf.
3. Investigate the root cause
This is where most CAPA processes fail. A survey of over 300 quality professionals found that poor root cause analysis was the single most cited problem with their CAPA systems.
The failure mode is stopping too early. "The operator didn't follow the procedure" is not a root cause. Why didn't they follow it? Was the procedure unclear? Was it physically impossible to follow given the setup? Were they never trained on it? Did the procedure not account for this material type?
Two common methods:
5 Whys. Start with the problem and ask "why" repeatedly until you reach a cause you can actually control. The rejected bore diameter: Why were parts oversize? The tool was worn. Why was the tool worn? It wasn't replaced at the scheduled interval. Why not? The schedule wasn't posted at the machine. Why not? There's no standard process for posting tool change schedules. Now you have something actionable.
Fishbone diagram. Organize potential causes into categories: machine, method, material, manpower, measurement, and environment. Work through each category to identify which factors contributed. This works well when the cause isn't obvious or when multiple factors combined to create the problem.
The goal is a root cause statement that, if addressed, would prevent the problem from recurring. If your root cause could apply to almost any quality problem ("lack of training," "human error"), you probably haven't dug deep enough.
4. Plan and implement corrective action
Design a fix that addresses the root cause, not the symptom. If the root cause is that tool change schedules aren't posted at machines, the corrective action isn't "retrain the operator." It's "create a standard process for posting tool change schedules at each work center and verify it's being followed."
Good corrective actions are specific. They name who is responsible, what they need to do, and when it needs to be done. "Improve the process" isn't a corrective action. "Engineering to update SOP-4520 to include tool change schedule posting requirements by April 15, and shift leads to verify compliance daily for 30 days" is one.
5. Implement preventive action
This is the step most manufacturers skip. You've fixed the problem that happened. Now ask: where else could this same type of failure occur?
If the root cause was missing tool change schedules at one work center, check whether other work centers have the same gap. If a supplier's material caused a nonconformance, review whether your incoming inspection process catches that type of defect for other materials.
Preventive action takes the lesson from one failure and applies it across the operation before the same failure shows up somewhere else.
6. Verify effectiveness
A corrective action isn't complete when it's implemented. It's complete when you can prove it worked.
Set a verification period. Define what success looks like in measurable terms. If the problem was oversize bore diameters, check the inspection data for that operation over the next 30, 60, or 90 days. If the defect rate dropped to zero or near-zero, the corrective action was effective. If the problem recurs, your root cause analysis missed something, and you need to reopen the investigation.
This step is the second most common failure point in CAPA processes. Teams implement the fix, close the CAPA, and never check whether it actually worked. The problem recurs six months later, and the cycle restarts.
Common CAPA mistakes (and how to avoid them)
Treating every issue as a CAPA
Not every defect needs a formal investigation. If you open a CAPA for every minor deviation, your quality team drowns in paperwork and the serious issues don't get the attention they deserve.
Use risk-based criteria. A CAPA makes sense when a problem is recurring, when it affects product safety, when it results in a customer complaint, or when an audit identifies a systemic gap. A one-time cosmetic defect on a non-critical part probably doesn't need one. Some shops use a pre-CAPA monitoring window: log the issue, and if it recurs within a defined period, escalate to a formal CAPA.
Never triggering a CAPA at all
The opposite problem. Some shops treat CAPA as something that only happens during audits. Problems get fixed informally, the same issues keep coming back, and there's no record of what was investigated or what was done. This is worse than CAPA overuse because systemic issues never get addressed.
If you're seeing the same problem more than twice, it needs a formal investigation with documented root cause analysis and verified corrective action.
Shallow root cause analysis
"Operator error" is the most common root cause finding, and it's almost never the actual root cause. Blaming the operator closes the CAPA on paper but doesn't prevent the next operator from making the same mistake.
Push past the human error. Ask what about the process, the training, the tooling, or the environment allowed the error to happen. That's where the fix lives.
Skipping verification
Implementing a corrective action and closing the CAPA on the same day means you assumed the fix would work without testing it. Sometimes that assumption is correct. Often it isn't. Build a verification check into every CAPA, and don't close it until the data confirms the problem is resolved.
Running CAPA in isolation
When only the quality department knows about CAPAs, the investigations are limited to what quality can see. The operator who noticed the problem, the maintenance tech who adjusted the machine last week, the engineer who designed the process: they all have information that quality might not have.
Cross-functional involvement produces better root cause analysis. It also builds buy-in for the corrective actions, which matters more than most quality managers want to admit.
Measuring your CAPA process
A CAPA system that runs without metrics is a CAPA system you can't improve. Track these numbers to know whether your process is working.
Recurrence rate. How often do closed CAPAs reopen because the problem came back? This is the single best indicator of whether your root cause analysis is effective. If your recurrence rate is above 10-15%, your investigations aren't going deep enough.
Time to closure. How long does it take from opening a CAPA to verified closure? Long closure times mean corrective actions are stalling, either because responsibility is unclear, resources aren't available, or the investigation is stuck. Track the median, not the average, since a few complex CAPAs will skew the mean.
On-time closure rate. What percentage of CAPAs close by their target date? Chronically overdue CAPAs indicate either unrealistic timelines or insufficient prioritization.
CAPAs by source. Where are your CAPAs coming from? Customer complaints, internal inspections, audit findings, operator reports? The distribution tells you where your detection is strong and where problems are slipping through to customers.
Corrective vs. preventive ratio. If nearly all your CAPAs are corrective (reactive), you're always playing catch-up. A healthy system includes preventive actions that come from trend analysis, risk assessments, and process reviews, not just from failures.
Moving CAPA off spreadsheets
Most small shops start with a CAPA log in Excel. A row for each CAPA, columns for the problem description, root cause, corrective action, responsible person, due date, and status. For a shop that processes five CAPAs a year, this is fine.
It stops working when volume increases, when multiple people need to update the same records, or when an auditor asks for the history of changes to a specific CAPA.
The problems are the same ones that push manufacturers from spreadsheets to dedicated software in other areas:
No audit trail. Anyone can edit any cell. If someone changes a root cause finding or a due date, there's no record of what it was before.
No workflow. There's no automatic routing from investigation to approval to implementation. CAPAs stall because no one is prompted to act on the next step.
No connection to production data. Your CAPA log is in one spreadsheet. Your inspection data is in another. Your nonconformance reports are in a folder somewhere. Connecting a CAPA to the actual quality data that triggered it means manual searching and copy-pasting.
A manufacturing ERP with quality management built in solves this by keeping CAPA records connected to the production data they reference. The nonconformance, the lot number, the inspection results, and the corrective action all live in the same system. When an auditor asks to see the history of a specific quality event, you pull it up in seconds instead of assembling it from three different files.
How to get started
If you don't have a formal CAPA process today, you don't need to build a full quality management system before you start. Here's a practical path.
Step 1: Pick your top three recurring quality problems. Look at your scrap reports, rework logs, and customer complaints from the last six months. Which problems keep showing up? Start your CAPA process with those. Solving three real, recurring problems is worth more than writing a procedure that covers everything but solves nothing.
Step 2: Investigate one of them properly. Take the most expensive recurring problem and run a real root cause analysis. Get the operator, the engineer, and the quality person in a room. Use the 5 Whys or a fishbone diagram. Write down the root cause, design a corrective action, assign an owner, and set a due date.
Step 3: Verify and document. After the corrective action is implemented, track the results for 60-90 days. Did the problem stop? Document what you found, what you did, and what the results were. This becomes your first completed CAPA record.
Step 4: Create a simple procedure. Based on what you learned from that first CAPA, write a one-page procedure that defines when to open a CAPA, who's involved, what the steps are, and how verification works. Keep it short. A ten-page CAPA procedure that nobody reads is worse than a one-page procedure that everyone follows.
Step 5: Build the habit. Run your CAPA process consistently for three months. Review open CAPAs in your weekly production meetings. Track the metrics. Adjust the procedure based on what's working and what isn't. The goal is a process that your team uses because it helps them, not one they endure because the auditor requires it.
Conclusion
CAPA is the difference between a shop that fixes the same problems over and over and a shop that fixes them once. Corrective action finds the root cause and eliminates it. Preventive action applies that lesson before the next failure occurs. Verification confirms the fix actually worked.
You don't need a dedicated quality department or an enterprise QMS to start. You need a willingness to investigate problems beyond the surface, the discipline to follow through on corrective actions, and a way to check whether those actions made a difference.
The manufacturers who run CAPA well aren't the ones with the thickest quality manuals. They're the ones whose operators and engineers know that when a problem gets investigated, it gets investigated properly, and it doesn't come back.
Ready to connect quality data to your production floor?
WorkCell gives manufacturers built-in quality tracking that links nonconformances, corrective actions, and inspection results to your actual jobs and materials. No separate QMS software required.
Book a demo and we'll show you how it works with your production data.