Voice of Customer (VOC)
Voice of Customer (VOC) captures raw customer feedback and translates it into structured, measurable quality requirements. ProBeya provides a complete VOC-to-CTQ workflow: from organizing customer statements through building Critical-to-Quality trees with specification limits, all the way to future SPC chart linkage.
VOC is the starting point of the Lean Six Sigma Define phase. Without understanding what the customer truly needs, every subsequent improvement effort risks solving the wrong problem.
Overview
Manufacturing and pharma teams often struggle to connect customer complaints, audit findings, and market feedback to actionable process improvements. Customer statements are qualitative ("the product feels inconsistent"), but process controls require quantitative specifications (viscosity between 45-55 cP at 25 deg C).
ProBeya's VOC module bridges this gap with a structured workflow:
- Collect raw customer statements (verbatim or summarized)
- Translate each statement into a structured need
- Decompose needs into a CTQ tree with measurable specifications (target, LSL, USL)
- Link CTQ specifications to SPC control charts (future integration)
Getting Started
- Navigate to a workspace and select Lean > VOC from the sidebar.
- Click New VOC Study and provide a title, description, and collection method.
- Define the customer segment being studied (e.g., "End users," "Distributors," "Internal QA").
- Start adding VOC items -- individual customer statements with their translations.
- Once enough voice data is collected, build a CTQ tree that decomposes high-level needs into measurable characteristics.
Key Concepts
| Term | Definition |
|---|---|
| VOC Study | A scoped investigation that captures and organizes customer feedback for a specific segment or topic. |
| VOC Item | A single customer statement (raw voice) paired with its team-interpreted need (translated need). |
| Raw Voice | The verbatim or summarized customer statement exactly as received. |
| Translated Need | The team's interpretation of the raw voice as a structured, actionable requirement. |
| CTQ Tree | A hierarchical decomposition: Need > Driver > Critical-to-Quality characteristic > Specification. |
| Specification Limits | The measurable bounds for a CTQ: Target value, Lower Specification Limit (LSL), and Upper Specification Limit (USL). |
| Collection Method | How feedback was gathered: surveys, interviews, focus groups, complaint logs, audit findings, etc. |
How It Works
VOC Study Lifecycle
Each VOC study progresses through a defined lifecycle:
| Status | Description |
|---|---|
| Draft | Initial creation. Team is planning the study. |
| Collecting | Actively gathering customer feedback through the defined collection method. |
| Analyzing | Feedback collection is complete. Team is translating voices and building CTQ trees. |
| Completed | Analysis is done. CTQ specifications are finalized and ready for process control linkage. |
Managing VOC Items
Each VOC item captures both the original customer signal and the team's interpretation:
- Raw Voice -- The exact customer statement or a faithful summary. Preserving the original language prevents interpretation drift.
- Translated Need -- The team's structured interpretation. For example, raw voice "The cap is hard to open" becomes translated need "Cap torque must be within ergonomic range for target demographic."
- Category -- Optional grouping (e.g., "Quality," "Usability," "Delivery," "Safety").
- Priority -- Low, medium, or high. Reflects the business importance of addressing this need.
- Frequency -- How often this statement was heard. Higher frequency signals stronger customer consensus.
- Source -- Where the feedback came from (e.g., "Q2 Customer Survey," "FDA Audit Finding #3").
Building CTQ Trees
CTQ trees decompose high-level customer needs into measurable quality characteristics. The tree hierarchy is:
Need (from VOC items)
└── Driver (what influences the need)
└── CTQ Characteristic (what to measure)
└── Specification (target, LSL, USL)
For example:
Need: Product consistency
└── Driver: Viscosity uniformity
└── CTQ: Viscosity at 25C
└── Spec: Target=50cP, LSL=45cP, USL=55cP
The tree structure is stored as JSONB because tree depth and branching vary per study. The entire tree is edited and saved as a unit in the UI, enabling fluid drag-and-drop restructuring.
SPC Integration (Placeholder)
The linkCtqToSpc procedure prepares the connection between CTQ leaf nodes and future SPC control charts. When the SPC module becomes available, each CTQ specification will link directly to a control chart that monitors whether the process stays within specification limits.
Currently, the link is stored as metadata within the CTQ tree's JSONB treeData field. This lightweight approach avoids creating a separate link table for a placeholder feature while preserving the data structure for future upgrade.
Configuration
| Setting | Description |
|---|---|
| Workspace | Every VOC study belongs to a workspace. |
| Customer Segment | Optional label identifying which customer group is being studied. |
| Collection Method | Describes how feedback is gathered (surveys, interviews, etc.). |
| Status | Controls the study lifecycle (draft, collecting, analyzing, completed). |
Permissions
| Role | Capabilities |
|---|---|
| Workspace Member | View studies, add/edit VOC items, create and edit CTQ trees |
| Workspace Admin | All member capabilities plus create/delete studies and manage study status |
| Organization Admin | Full access across all workspaces |
All operations enforce multi-tenant isolation. VOC items and CTQ trees carry a denormalized organizationId from their parent study to enable direct tenant-scoped queries without joins.
Tips & Best Practices
Always record the customer's actual words before translating them. The raw voice is your primary evidence. If the translation later proves inaccurate, the raw voice lets the team course-correct without going back to the customer.
When multiple customers mention the same need, track the frequency. A need mentioned by 80% of respondents should take priority over one mentioned by 5%, regardless of how eloquently it was expressed.
CTQ decomposition requires input from quality engineers, process owners, and customer-facing staff. A tree built in isolation by one function will miss critical drivers or set unrealistic specifications.
Each CTQ specification should map to a measurable process parameter that operators can actually control and monitor. If a specification cannot be measured on the shop floor, it is not actionable.
Troubleshooting
| Issue | Resolution |
|---|---|
| Cannot add VOC items | Verify the parent study exists and belongs to your organization. The system validates study ownership before accepting new items. |
| CTQ tree changes not saving | The entire tree is replaced on each save. Ensure you are saving the complete tree structure, not a partial update. |
| SPC link returns placeholder message | The SPC module is not yet implemented. The link is stored for future use and will become functional when SPC charts are available. |
| Study not appearing in list | Studies are workspace-scoped. Confirm you are viewing the correct workspace. Use the status filter to ensure the study's current status matches your filter criteria. |
Related Features
- SIPOC Diagrams -- Define process boundaries that frame the scope of VOC studies.
- FMEA Analysis -- Failure modes with the highest severity are often directly related to unmet CTQ specifications.
- Standard Operating Procedures -- SOPs should include controls that keep processes within CTQ specification limits.
- Kaizen / Problem Solving -- When a CTQ specification is chronically missed, initiate a Kaizen project to investigate root causes.
- KPI Boards -- CTQ metrics can be monitored as KPIs on operational boards.