Satisfaction, NPS, and Complaint Analysis — Turning Customer Voices into Actionable Improvement Items

Abstract: The annual customer satisfaction survey scored 87 points, and management was satisfied—however, during the same period, complaints increased by 23%, and NPS fell from +12 to +3. Focusing solely on satisfaction can lead to "self-comfort"; NPS looks at the willingness to recommend, while complaints focus on failure events. This article explains the differences and combined use of CSAT, NPS, and complaint data, the structured coding and root cause closure of complaints, the conversion path from VOC to CTQ, and a monthly quality meeting template for B2B manufacturing scenarios—forming a complete closed loop with 1.3.1 VOC and Customer Journey.

1. Case Study: Good Scores, but Losing Customers

A B2B industrial equipment company sends out a satisfaction questionnaire (5-point scale) every six months, with an average score of 4.2/5, and 82% of respondents indicating "satisfied." The sales department reported "good customer relationships."

Meanwhile:

• NPS (How likely are you to recommend us? 0-10): Promoters (9-10) decreased from 45% to 28%, Detractors (0-6) increased from 12% to 31% → NPS = 28% − 31% = −3
• Formal Complaints (written/8D triggered): increased from an average of 18 per year to 41
• Lost 2 strategic customers, with the reasons for exit written in competitor bids: "slow response, repeated issues not closed"—these reasons did not appear in the open-ended questions of the satisfaction survey (customers were unwilling to write long responses).

The CEO asked the quality director: "Our satisfaction is 4.2, why are we still losing orders?"

Root Cause: The satisfaction survey had sample bias (only sent to friendly procurement contacts), questions were too broad ("overall satisfaction"), and lack of integrated analysis with complaints/NPS—the three types of data were not aligned.

2. Three Types of Indicators: What Each Measures

Principles for Combined Use:

• CSAT: Look at trends and segments (product/region/customer manager)
• NPS: Look at loyalty and risk (focus on Detractor interviews)
• Complaints: Look at systemic defects (must be closed-loop)

All three types of data should be integrated into the VOC database (1.3.1) and then mapped to CTQ.

3. NPS Calculation and Interpretation

NPS = Promoter Percentage − Detractor Percentage (Passives 7-8 are not included in the calculation but should be analyzed)

Example: 200 valid responses

• Promoters (9-10): 56 people → 28%
• Passives (7-8): 82 people → 41%
• Detractors (0-6): 62 people → 31%
• NPS = 28% − 31% = −3

B2B Considerations:

• Multiple contacts from one customer should be deduplicated or stratified (technical vs. procurement)
• NPS fluctuates more with small sample sizes (<30)—look at rolling quarters

Actions:

• Detractors: Executive follow-up within 30 days, root cause projects
• Passives: Ask "What would it take to get a 9?"
• Promoters: Use as case studies, co-publish (with consent)

4. Complaint Management: From Events to Data

4.1 Definition and Channels of Complaints

Formal Complaints: Customer written/email/portal registration, triggering recording, response SLA, root cause analysis, 8D.

Must be distinguished from "general inquiries" and "logistics damage"—classification rules should be documented.

4.2 Minimum Data Fields

Use the same failure coding system as 10.2.1 Warranty and Claims Analysis.

4.3 Monthly Complaint Analysis Meeting

Inputs: Current month's complaint list, NPS Detractor list, low CSAT items

Outputs:

1. Top 3 failure mode Pareto
2. List of repeated complaints (must be escalated)
3. New improvement projects (Owner, date)
4. Confirmation of closed complaint customers

Prohibited: Complaint closure = writing a report without verifying customer satisfaction.

5. Case Study: From Complaint to CTQ

Complaint: Customer's production line stopped for 4 hours due to a connector having excessive insertion force, batch 20260412.

Analysis Chain:

1. Failure Mode: Insertion force exceeds specification upper limit
2. Traceability: Injection mold wear + no first article insertion force testing
3. CTQ: Insertion force 35～45 N (customer CSR)
4. Process Control: Increase detection in PFMEA; add first article and 2-hour sampling inspections per shift in CP
5. Supplier/Internal: Change mold maintenance cycle from 50,000 to 30,000 cycles

Verification: No complaints for the same failure mode in the following 3 months; the customer's NPS increased from 6 to 8.

6. Key Points for Designing Satisfaction Questionnaires

Avoid:

• Only "overall satisfaction"
• Sending only to procurement, not to user departments
• Annual surveys—key customers should have quarterly pulse checks

Recommended Dimensions (B2B manufacturing):

• Quality and consistency
• Delivery and response
• Technical/support
• Change management communication
• Problem closure speed

Each question should be scored 1～5 + "most needed improvement" open-ended question (open-ended questions must be 100% classified and coded, not just stored in Excel without review).

7. Interface with ISO 9001 / IATF

• 9.1.2 Customer Satisfaction — Must monitor customer perceptions; all three types of data can be used as inputs
• 10.2 Nonconforming Products and Corrective Actions — Complaints are a primary entry point
• Management Review — Complaint trends, NPS, and CSAT should be used as inputs

IATF emphasizes customer-specific requirements (CSR)—CSR-related items in complaints should be separately tracked.

8. VOC to Improvement Closure (Integration with 1.3.1)

CSAT/NPS/Complaints → VOC database classification → priority (severity × frequency)
    → CTQ/specification updates → PFMEA/CP/training → verification → customer feedback

Priority Matrix:

9. Implementation Roadmap (120 Days)

0～30 Days: Standardize complaint fields; create a failure mode coding table; publish SLA

31～60 Days: First NPS pilot (TOP 20 customers); revamp CSAT questionnaire

61～90 Days: Monthly complaint + VOC meetings; conduct in-depth interviews with 3 Detractors

91～120 Days: First quarterly management review report on the three indicators; close 2 CTQ improvement projects

10. Case Study 2: CSAT Segmentation Reveals "Hidden Dissatisfaction"

150 CSAT responses (5-point scale) from the same customer group, overall average 4.1—seems healthy.

Segmentation by Dimension:

Conclusion: The overall 4.1 is skewed by "Product Quality"; response and change communication are the leading indicators for NPS decline and complaint increase.

Actions: Include "Issue Response" in sales and quality KPIs—first response ≤4 hours, initial 8D draft ≤5 working days; next quarter's CSAT target for this dimension ≥3.8.

Lesson: Never look at only the total score—B2B procurement may give 4 points because "there are no other choices," while user departments may rate 5 (Detractor) in NPS.

11. Silent Customers and Attrition Warnings

A significant proportion of B2B attrition involves customers who do not complain, do not give low scores, but simply do not renew orders.

Warning Signals:

Countermeasures: Conduct quarterly Business Reviews with strategic customers (not just satisfaction questionnaires); mandatory executive visits for Detractors + declining orders.

Consistent with monitoring "use/service" touchpoints in the 1.3.1 VOC customer journey.

12. Template for Monthly Quality Meetings (60 Minutes)

Participants: Representatives from quality, sales, engineering, and production—avoid "quality department looking at complaints in isolation."

13. Implementation Checklist

Data Foundation:

• Unified failure coding for complaints/warranty/8D
• CSAT questionnaire with ≥5 dimensions + open-ended question coding process
• Documented NPS survey method (who fills out, frequency, sample size)

Analysis and Closure:

• Monthly dashboard for the three indicators (CSAT/NPS/Complaints)
• 100% follow-up rate for Detractors within 30 days
• Complaint closure must be confirmed by the customer or equivalent verification
• Top complaint items mapped to PFMEA/CP update records

Management Review:

• Quarterly report on complaint cost trends
• List of failure modes not covered by FMEA

14. Common Pitfalls

15. Summary

Satisfaction tells you "how they feel," NPS predicts "whether they will continue to cooperate," and complaints tell you "what is already broken."

Using all three metrics together, the customer and market segment can have warnings + root causes + improvements—not just a pretty PPT at year-end.

Next steps: Create a Pareto chart of failure modes for the past 12 months and compare it with DFMEA—the top 3 failure modes not covered by FMEA are the entry points for integrating VOC into the system.

Knowledge code: 1.3.2

Version: v20260528

Author: Quality Think Tank