I. The four-minute changeover
NarratorThe fixture has been installed for a week. Today the project engineer is closing out one more requirement before the FAT report goes to the customer.
- Engineer“The requirement says changeover must take five minutes or less. We timed it. Four minutes.”
- Shift supervisor“Who changed it over?”
- Engineer“Our two technicians.”
- Shift supervisor“I have one operator on nights. Can she do it on her own?”
What did the four-minute test prove? What would you want to see before signing off?
What the stopwatch proved, and what it didn’t
The stopwatch gives us verification evidence for the time limit, under the conditions tested. The supervisor has raised a validation question about how the fixture will really be used.
NarratorNow listen to one more exchange. This is the one most teams never have.
- Shift supervisor“When did you stop the clock?”
- Engineer“When the last clamp was tight.”
- Shift supervisor“On my line, changeover ends at the first good part. First-piece check takes us three minutes.”
So the four minutes hides three separate gaps. Only one of them is about validation.
- 1Who and with what. Two technicians and a full cart, not one operator with the tools at the machine. During the test, one technician held the 14 kg plate while the other started the clamps. This is the validation gap.
- 2Where the clock starts and stops. The requirement never said. Clamps tight is not first good part. This is a verification gap: there was no agreed pass definition, so the evidence is weak even for the requirement as written.
- 3How many times. One run by people who installed the fixture. A single time is a demonstration, not a capability. Also a verification gap.
Invite a night-shift operator to try a representative changeover, alone, with the tools available on that shift, from last good part to first good part. Watch, time it, and note every moment she stops, reaches, or asks for help.
Rewrite the requirement so one test answers both questions
When the requirement already describes the real user, the real task and the real conditions, a single trial is both verification and validation. Same test, both words, both correct. That is why experienced people sometimes use the two words for one activity.
Write these five parts into the functional specification before design review. Then the engineer knows what counts as a pass, the project manager knows who must try it, and QC knows exactly what a signature covers.
Standards note. ISO 9000:2015 defines verification as confirmation, by objective evidence, that specified requirements are met, and validation as the same confirmation for a specific intended use or application. ISO 9001:2015 clause 8.3.4 asks for both as separate design and development controls. The last-good-part to first-good-part definition of changeover comes from SMED practice.
II. Call it: verification, validation, or both?
Many of us learned a quiet shortcut: small, early, paper checks are verification; big, late, live tests are validation. It feels right, and it is wrong often enough to cause real trouble.
The question you are answering tells you which activity you are doing. Not the size of the test, not the stage of the project, and not who holds the clipboard.
Ten situations from real projects. Make the call, then read why.
1. A designer checks the electrical drawings against the approved functional specification.
Small, early, desk work, and still verification: does this output match what we specified?
2. At FAT, the full line runs 1,000 cycles at 12 parts per minute, as the specification requires.
Big, late and live, and still verification. It checks a specified rate. Nobody’s working day was tested.
3. An operator holds a cardboard mock-up of the new control station to check reach and height.
Small, early and rough, and still validation: can this person do the job with it?
4. An electrician checks every I/O point against the I/O list during commissioning.
Every point against a specified list. Classic verification.
5. A night operator reads an alarm on the HMI and cannot tell which guard door is open.
The alarm may match the alarm list exactly. The question here is whether a real user can act on it at 2 a.m.
6. QC does a first-article inspection on the CMM against the drawing.
Measured features against specified dimensions. Verification, however thorough.
7. The specification says a maintenance technician wearing work gloves must replace a photo-eye in under ten minutes. A technician from the plant tries it.
The requirement already names the user and conditions, so one trial answers both questions.
8. A supervisor runs the new work instruction with a new hire on day one.
Can the intended user get the job done with the document? That is validation, even for paperwork.
9. The customer insists on a run-off with their real part mix, including the warped parts they actually receive, and the spec lists that mix.
Real parts, real conditions, written into the spec. Both. Without the spec line, it would be validation only, and any failure becomes a commercial discussion.
10. A safety function is tested to confirm it reaches the performance level stated in the safety requirements specification. The standard calls this “validation”.
A trap on purpose. ISO 13849-2 uses the word validation, but by our two questions this is mostly checking against a specification. Use the standard’s word in the document. Know which question you are answering.
III. The stainless-steel tray
NarratorForty trays are on order. The first one has arrived from the fabricator. Priya in QC measures it before the other 39 are released.
| Feature | Drawing | Tolerance | Measured | Result |
|---|---|---|---|---|
| Length × width | 600 × 400 mm | ±1.0 mm | 600.4 × 399.7 mm | Pass |
| Depth | 80 mm | ±1.0 mm | 79.6 mm | Pass |
| Handle length × clearance | 110 × 35 mm | ±0.5 mm | 109.6 × 34.8 mm | Pass |
| Handle bends | R5 | — | R5 gauge fits | Pass |
| Edges | Deburred | — | No burrs found | Pass |
Priya is asked to sign off the first article and release the other 39 trays. What should she sign today?
NarratorMarco from first shift tries the tray with a load of wet parts, about 9 kg. He is wearing the required cut-resistant gloves with gauntlet cuffs.
- Marco“I can get two fingers under the handle, to the first knuckle. Lifting is okay. But when I tip it to empty it, it starts to slide.”
- Priya“It’s exactly to drawing.”
- Marco“I believe you. The drawing just never wore these gloves.”
Verification question
Do the tray dimensions and handle clearance match the approved drawing?
Yes. The readings support it.
Validation question
Can the operator grip, lift and empty the loaded tray while wearing the required gloves?
Not yet. The tipping step fails.
What finding it later would have cost
The same gap, found at four different moments. Figures are an example for 40 trays; your numbers will differ, but the shape will not.
Validation did not need a finished tray. It needed one person, the real gloves and a rough sample, before the drawing was approved.
Tool: the validation trial card
Use it for a mock-up, a first article, a FAT visit or a pilot run. Agree the pass before anyone tries it.
- Who
- A real user from the shift that will use it. Not the designer, not the technician who built it.
- Task
- The whole job, start to end. Lift, carry, tip, empty, return. Not just “hold it”.
- Conditions
- Real PPE, real load, real tools, real lighting, real time pressure.
- Watch for
- Hesitation, workarounds, asking for help, two hands where one was planned. Watch; do not coach.
- Pass
- Agreed before the trial, in plain words: “empties a 9 kg load with gloves, without the tray slipping, three times in a row.”
Tool: the sign-off statement
For QC. A signature should say exactly what the evidence covers. Edit the fields; the statement updates. The tray example is filled in.
Standards note. ISO 9241-11:2018 describes usability as effectiveness, efficiency and satisfaction for specified users in a specified context of use; the gloves are part of that context. For loaded lifting, ISO 11228-1 covers manual handling limits. ISO/IEC 17025 practice is a good model for stating instrument, conditions and scope with every result.
IV. The Excel inspection report
NarratorDaniel, a quality engineer, has built a workbook that flags failed bore readings. The rule is simple: diameter 12.00 millimeters, plus or minus 0.05. Before releasing it, he tests it with values he already knows the answer to.
| Test value (mm) | Why this value | Expected | Workbook |
|---|---|---|---|
| 12.000 | Nominal | Pass | Pass |
| 12.050 | Upper limit, exactly | Pass | Pass |
| 12.051 | Just above upper limit | Fail | Fail |
| 11.950 | Lower limit, exactly | Pass | Pass |
| 11.949 | Just below lower limit | Fail | Fail |
| Blank | Missing reading | Missing | Missing |
| “12,02” | Comma decimal from an exported file | Error | Error |
Keep this set. Nominal, each limit, just outside each limit, a blank and a wrong format: these seven values verify almost any pass/fail rule, in a spreadsheet, a PLC or a vision system.
Daniel says the workbook is ready. What has he shown?
NarratorFriday, 3:35 p.m. The truck leaves at 5. Grace, the second-shift supervisor, opens the report.
- Grace“I see four red readings on Machine 2. Which lot do I hold?”
- Daniel (on the phone)“The lot numbers are on the second tab.”
- Grace“The second tab shows the lot changed at 1:40. Two of the reds are before that, two after. So is it 2611-A, 2611-B, or both?”
She holds 2611-B, because it is the lot on the screen. Lot 2611-A is already on the truck. On Monday, the customer’s receiving inspection finds it.
Verification question
When we enter known pass, fail and boundary values, does the workbook flag each reading according to the agreed rules?
Yes. Every test value was flagged correctly.
Validation question
Can the supervisor use the report to identify the affected lot and make the correct hold decision before shipment?
No. The flags were right; the decision was wrong.
Daniel built the report around the rule. Grace needed it built around the decision. The fix is small: show the lot on every reading and sort by lot. Finding that out was the expensive part.
Give a supervisor a realistic data set with a lot change in the middle, and ten minutes to make the hold call. Watch where she looks. That is the whole validation trial.
Standards note. ISO 9001:2015 clause 8.7 requires control of nonconforming outputs; the hold decision is that control. In regulated industries, spreadsheets used for quality decisions are often treated as computerized systems (for example under GAMP 5), where intended-use testing is expected as well as functional testing.
V. When the gap shows up at SAT
- Plant manager“Your FAT report says four minutes. Last night my operator took eleven, and she had to wake up maintenance to help lift the plate.”
- Project manager“We met the five-minute requirement at FAT.”
- Plant manager“It doesn’t work for us. I’m not signing SAT until it does.”
Is this a defect to fix under warranty, or a change?
Many integration contracts tie acceptance to the agreed specification. That makes verification your commercial protection, and the record of test conditions decides how strong it is.
If the FAT record says
“Changeover: 4 min 05 s. Two technicians, full tool cart, clamps tight. One trial.”
The gap is visible. One operator was never specified or tested. This is a requirement gap to resolve together, often as a change.
If the FAT record says
“Changeover: 4 min. Pass.”
It reads like a promise. The customer hears “it works”, and the gap lands on you as a warranty claim or an unpaid service call.
Three moves that protect FAT, SAT and the warranty
- 1Put users and conditions in the functional specification and get it signed: who operates, which shift, which PPE, which tools, where the clock starts and stops.
- 2Invite a customer operator to FAT. One day of their time is the cheapest validation you will ever buy, and the gap shows up in your shop, not theirs.
- 3Record the test conditions on every FAT and SAT line. A result without conditions is a promise you did not mean to make.
Note. Who pays in a real dispute depends on the contract wording and the documents both sides signed. Bring the specification, the FAT record and the change log to that conversation.
VI-1. A line you can use tomorrow
Each role gets one sentence to say out loud in the next meeting. It works because it makes the room answer one of the two questions.
- Engineer“Which requirement are we checking, and what result would count as a pass?”
- Project manager“Whose working day are we designing for, and have we seen that person try it?”
- Quality control professional“What can I confidently sign off from this evidence, and what is still open?”
Phrases that should make you pause
You will hear these in design reviews, FAT meetings and Friday afternoons. Each one hides a mix-up between the two questions.
| You hear | What is mixed up | What you can say |
|---|---|---|
| “It passed FAT, so it works.” | Verification taken as validation | “It met the spec under FAT conditions. Who has tried it the way the plant will run it?” |
| “The customer will validate at SAT.” | Validation pushed to the most expensive moment | “What can we let an operator try before it ships?” |
| “We validated the drawing.” | A drawing check is verification | “Against which requirement? And has a user seen a mock-up?” |
| “Operators will get used to it.” | An open validation gap with no owner | “Let’s watch one try it and decide if that is acceptable.” |
| “It’s within tolerance.” | True, and maybe not enough | “Agreed. Is it the right tolerance for the job?” |
VI-2. Back to the night shift
NarratorWe started here. Let’s finish here.
The two technicians finished in four minutes. What can you sign off today, and what would you test next?
Today you can sign off: “Changeover completed in 4 min 05 s, clamps tight, by two technicians with the full tool cart, one trial.” That is real verification evidence, stated honestly.
Still open: one operator alone, with the tools at the machine, to first good part, repeatably.
Test next: a night-shift operator does three representative changeovers on her own, last good part to first good part. Record the time, any help she needed, and any tool she had to go and find. If the requirement should have said “one operator”, raise it through change control now, before FAT, while it is still cheap.
Next time someone asks, “Which one are we doing right now?”, you won’t need to pause. Ask what question the test is answering. The answer will tell you.
VII. Practice checks
Five quick checks. Pick an answer to see the explanation. You can change your answer at any time; practice does not count toward the certificate.
1. Running the full production line for eight hours at FAT is always validation.
If the run checks a specified rate, uptime or reject limit, it is verification. The question decides, not the size of the test.
2. If a requirement does not say where the clock starts and stops, a timed test gives weak verification evidence.
Without an agreed pass definition, two honest people can time the same changeover and get four minutes and seven.
3. A tray made exactly to drawing has been validated.
Made to drawing is verification. Whether an operator can use it with gloves is a separate question.
4. Inviting a customer operator to FAT can give early validation evidence.
A real user trying the real task in your shop finds the gap before SAT, when it is still cheap to fix.
5. A good sign-off says what the evidence does not cover.
Stating what is still open protects the signer, the project and the customer. A bare “Pass” reads like a promise.
VIII. The label printer at the cold dock
One scenario, seven decisions. You will see the explanations after you submit. Six correct out of seven passes. If you do not pass, you can review the explanations and try again.
The requirement
“The print-and-apply unit shall print and apply a GS1-128 shipping label to each case within 6.0 seconds. Labels shall be scannable.”
The FAT record
200 cases run at a 6.0 s cycle. Every label printed and applied. Labels scanned at the bench with a handheld scanner: 200 of 200 read. Room temperature 21 °C. Cardboard cases supplied by the integrator. Label data checked against the customer’s GS1 label specification, Rev C, on 10 sampled labels: all correct.
What you learn at the pre-shipment meeting
At the customer’s dock, cases come straight from a −20 °C freezer into a 2 °C dock and carry condensation. Labels are read by a fixed overhead scanner about 2 m above the conveyor, at belt speed. The customer uses their own case stock.
Rate this course
Optional: how useful was this course for your work? Your rating helps us improve it. It is not needed for the certificate.
Thank you. Your rating is recorded.
Your certificate
Enter your legal name exactly as it should appear on the certificate. Once issued, the certificate ID never changes.
Your worksheet
Pick one requirement from your current project. Fill in both sides, then save it as a PDF or print it for your next design review or FAT meeting.
Which one are we doing right now?
PMCLEAR CPD / PDH worksheet
Verification: What did we specify, and what evidence shows we met it?
Validation: Can the intended users get their job done with this solution under the conditions they actually face?
Verification
Validation
Sign-off
Next action
About this training
Verification & Validation — 2-Hour CPD / PDH Course is a self-paced online professional development course by PMCLEAR for engineers, project managers and quality control professionals in Canada, the United States and internationally. It teaches how to tell, in a live meeting, whether a test or check is verification (did we meet what we specified?) or validation (can the intended users do their job under real conditions?), and how to sign off only what the evidence shows.
Learning outcomes
- Tell verification from validation by the question a test answers, not by its size or project stage.
- Write requirements that state the user, conditions, start and stop points, limit and number of trials, so one test can verify and validate.
- Design verification evidence: pass criteria, sample size, instruments, and boundary and wrong-format test values.
- Plan early, low-cost validation trials with real users under real working conditions.
- Write sign-off statements that say what the evidence covers and what remains open.
- Protect FAT, SAT and warranty positions by recording test conditions and putting intended-use conditions in the functional specification.
Course content (8 units)
- The four-minute changeover
Timed changeover evidence, missing pass definitions, single trials, requirement rewrite - Call it: verification, validation, or both?
Ten classification exercises from design review, FAT, commissioning and quality - The stainless-steel tray
First-article inspection, glove trial, cost of late discovery, validation trial card, sign-off statement - The Excel inspection report
Test-value design for pass/fail rules, decision-centered validation of reports - When the gap shows up at SAT
Specification, FAT records, change versus warranty, protecting acceptance - A line you can use tomorrow; back to the night shift
Meeting language for engineers, project managers and QC; closing decision - Practice checks
Five Yes / No practice checks with explanations - Final assessment
A seven-decision scenario: what can be signed off, what is open, what evidence is needed next
Format and completion requirements
- Narrated shop-floor stories with decision points, classification exercises, workplace tools (sign-off statement, validation trial card) and a workplace worksheet.
- To complete the course the learner had to complete every story decision and exercise, answer all practice checks, and pass the final scenario assessment (at least 6 of 7 decisions correct). Only then is the certificate issued.
Hours, subject type and where it counts
- Designed duration: 2 hours (120 minutes), 8 units. Subject type for every unit: engineering management and practice. The certificate needs at least 72 minutes of recorded active study time; it shows your own recorded time next to the designed duration.
- Canada: report it as CPD in your regulator's records; Canadian engineering regulators do not pre-approve providers or courses. British Columbia: continuing education hours.
- United States: PDH in most states. Not for PE license renewal in Florida, New York, New Jersey, Maryland or Indiana, which accept only board-approved providers.
- Ingénieurs du Québec : cours en anglais, 2 heures. Vous les déclarez comme formation continue. Ce n’est pas une formation de l’Ordre des ingénieurs du Québec.
Course duration: 2 hours. Delivery: self-paced online. Provider: PMCLEAR. Acceptance of CPD / PDH hours is decided by the learner's professional regulator or licensing board; this course is not accredited or pre-approved by an engineering regulator.