Contents
What is the NTF rate?
NTF (No Trouble Found) — also called CND (Can Not Duplicate) or RTOK (Re-Tested OK) — is the percentage of returned or rejected products that, upon re-inspection or re-test, show no detectable fault. The return is classified as NTF, the product is restocked or scrapped, and the cost is absorbed.
Something happened during production, handling, transport, or use that created a symptom. But by the time the product reaches the lab, the symptom is gone. That makes NTF one of the most expensive and most frustrating quality problems in discrete manufacturing — because the cost is real but the cause is invisible.
How do you calculate the NTF rate?
NTF Rate = (Number of NTF Returns / Total Number of Returns) × 100 %
Where:
- Number of NTF Returns = returned or rejected units for which the re-inspection or re-test found no fault. The product meets all specifications when tested.
- Total Number of Returns = all units returned under warranty or rejected by the customer, regardless of whether a fault was confirmed or not.
| Worked example — Connector assembly, Q1 2026 | Value |
|---|---|
| Total units shipped | 50,000 |
| Total customer returns | 120 |
| Returns with confirmed fault | 48 |
| Returns classified as NTF (no fault found) | 72 |
| NTF Rate (72 / 120 × 100 %) | 60 % |
An NTF rate of 60 % means that 6 out of 10 returned products are not confirmed as defective in the re-test. The return cost — logistics, re-inspection labour, restocking or scrapping, administrative processing, and customer relationship damage — is incurred without any corresponding product improvement, because there is no confirmed root cause to fix. In automotive and electronics manufacturing, NTF rates between 30 % and 70 % of all warranty returns are commonly reported in industry literature. Each NTF return typically costs between €50 and €500 in total handling cost, depending on the product complexity and the depth of re-inspection.
What are the 5 root causes of NTF returns?
NTF does not mean "nothing happened." It means the re-test cannot reproduce the problem. The fault existed — but it was intermittent, environmental, or destroyed during return handling. Here are the 5 actual root causes, ranked by frequency in discrete manufacturing:
| # | Root cause | What actually happened | Why re-test finds no fault | How MES data prevents it |
|---|---|---|---|---|
| 1 | Intermittent fault | A solder joint cracks under thermal cycling, a connector contact lifts under vibration, a seal leaks under specific pressure/temperature combinations. The fault appears only under certain operating conditions. | Lab re-test at room temperature and no vibration does not reproduce the field condition. The part passes all standard tests. | MES process data records the actual production parameters (soldering temperature, press force, curing time) for each unit. If the returned unit was produced during a process excursion (parameter near upper or lower spec limit), the MES data reveals the correlation — even when the re-test does not. |
| 2 | Test coverage gap | The end-of-line test does not cover the failure mode the customer experiences. Example: the test checks electrical continuity but not insertion force. The customer's assembly process requires a specific insertion force — which this unit does not meet. | Re-test uses the same end-of-line test protocol — which passed the part originally and passes it again. The failure mode is real but untested. | — |
| 3 | Handling or transport damage | The product was functional when shipped. Damage occurred during transport, storage, or customer-side handling before installation. ESD damage to electronics is a classic example — the component was damaged by electrostatic discharge at the customer's facility. | The damage is microscopic (ESD) or the product was further damaged during return shipping, overwriting the original fault signature. | — |
| 4 | Customer application error | The product is used outside its specified operating conditions. Wrong voltage, excessive temperature, incorrect mounting orientation, or incompatible mating part. The product "fails" in the customer's application but passes all tests within its specification. | Re-test confirms the product meets its specification. The mismatch is between the specification and the customer's application, not a defect. | MES data does not directly prevent this cause. |
| 5 | Inadequate re-test method | The fault is real and reproducible, but the re-test lab does not have the right equipment, environment, or procedure to detect it. The lab tests at 23 °C; the fault occurs at 85 °C. The lab tests static; the fault occurs under vibration. | The re-test method is insufficient, not the product. | MES process data for the returned unit shows which production parameters were at the edge of the process window. This guides the lab to test under those specific conditions — turning a blind re-test into a targeted investigation. |
The common thread: every NTF root cause is an information gap. The quality lab receives a product with no context — no production history, no process parameters, no test results, no alarm records.
How does an MES reduce the NTF rate?
An MES reduces NTF through mechanisms — each addressing a different part of the NTF problem:
| Mechanism | What the MES provides | NTF root cause it addresses | Practical example |
|---|---|---|---|
| Process parameter monitoring | Real-time capture of actual process parameters (temperature, force, time, speed) per cycle. | Intermittent faults, inadequate re-test method | The SYMESTIC process data module captures parameters per cycle from the PLC. |
How does NTF relate to other quality metrics?
| Metric | What it measures | Relationship to NTF |
|---|---|---|
| OEE Quality rate | First-pass yield at the end-of-line test | A high OEE Quality rate means few parts are rejected during production. But it says nothing about field returns. A plant can have 99.5 % first-pass yield and still have a 50 % NTF rate — because the end-of-line test does not cover all field failure modes. |
| PPM (Parts Per Million) defect rate | Total confirmed defects per million parts shipped | PPM counts only confirmed defects. NTF returns are excluded from PPM by definition (no fault found). A high NTF rate therefore masks the true field quality level: the customer experiences failures, but they do not appear in the PPM metric. |
| Warranty cost | Total cost of warranty claims per period | NTF returns are a direct component of warranty cost. |
| Customer satisfaction / 0-km reject rate | Rate at which the customer rejects parts at incoming inspection | NTF at incoming inspection is common in automotive Tier-1 supply chains. The customer's incoming test rejects a part, the supplier re-tests it and finds no fault. The dispute consumes engineering time on both sides and erodes the commercial relationship. |
FAQ
Is a 0 % NTF rate realistic?
No. Some level of NTF is inherent in any product with intermittent failure modes, environmental sensitivities, or complex customer applications. The goal is not zero NTF — it is to understand each NTF category and reduce the preventable ones. Intermittent solder joint failures can be reduced through tighter process control (MES process data). Test coverage gaps can be closed by correlating field complaints with tested parameters. Customer application errors can be reduced through better documentation. A realistic target for a mature NTF reduction programme is to reduce the NTF rate by 30–50 % from the baseline, not to eliminate it entirely.
Why is NTF particularly problematic in automotive?
In automotive supply chains, returned parts go through a structured 8D problem-solving process. NTF returns stall the 8D at step D4 (root cause analysis) because there is no confirmed defect to analyse. The customer expects a corrective action. The supplier cannot provide one — because the product passes every test. This creates a standoff that damages the business relationship.
How does NTF relate to Jidoka (built-in quality)?
Jidoka means building quality inspection into the production process so that defects are detected at the point of creation, not at end-of-line or in the field. NTF is the inverse signal: it represents defects (or perceived defects) that escaped every quality gate. Reducing NTF by improving in-process detection — e.g., monitoring a critical process parameter in real time and stopping the machine when the parameter drifts — is a direct application of the Jidoka principle. The MES process data module enables Jidoka digitally: parameter thresholds trigger alerts before a borderline unit reaches the end of the line.
What is the difference between NTF, CND and RTOK?
They describe the same phenomenon with different terminology depending on the industry and company. NTF (No Trouble Found) is the most widely used term across industries. CND (Can Not Duplicate) is common in electronics — it emphasises that the reported symptom could not be reproduced in the lab. RTOK (Re-Tested OK) is common in automotive supply chains — it emphasises that the returned unit passed re-testing. The underlying problem is identical: the customer reports a fault, the supplier cannot confirm it, and the root cause remains unknown.
Related: Quality Control · Jidoka · Poka-Yoke · OEE Explained · SYMESTIC Process Data · SYMESTIC Alarms Module · SYMESTIC Production Metrics · MES: Definition & Functions
Calculate what defective parts cost you
In a free 30-minute call we go through scrap and the capacity lost to retests and calculate the annual cost with your own numbers.