Customer Data Collection

Customer Data Collection (CSSGB BoK II.B.2 — Apply)

Quick Answer: Collect the Voice of the Customer with surveys, focus groups, interviews, and observation. Choose methods for the insight you need, then write questions that are specific, single-purpose, and neutral—free of vagueness, ambiguity, and bias—so VOC can become measurable CTQs.

Once you know who the customers are (II.B.1), you must hear them accurately. BoK II.B.2 is Apply: select methods and critique or improve question quality.

VOC Data: Reactive vs Proactive

Best practice blends reactive signals (what is already broken) with proactive VOC (structured listening) and process data.

Core Collection Methods

Surveys

Surveys gather structured responses from many people via web, email, phone, or paper.

• Strengths: breadth, quantification, trend tracking, segmentation
• Weaknesses: low response bias, shallow “why,” poor design yields misleading scores
• Green Belt use: baseline satisfaction, prioritize themes, track post-improve scores

Use Likert scales, rankings, or closed choices for analysis; include a few open comments for language—not 40 open essays no one will code.

Focus groups

Focus groups are moderated small-group discussions (often 6–10 participants).

• Strengths: customer language, interaction that surfaces issues individuals forget, rapid theme generation
• Weaknesses: groupthink, dominant voices, not statistically representative
• Green Belt use: explore needs early; clarify survey wording; probe emotional drivers (Kano “delight” clues)

Interviews

Interviews are one-on-one (or one-on-two) conversations, structured, semi-structured, or open.

• Strengths: depth, follow-up probes, confidential candor, B2B multi-role listening
• Weaknesses: time cost, interviewer bias, small samples
• Green Belt use: critical external accounts, internal customer handoff pain, CTQ drafting from stories

Observation (including Gemba)

Observation watches real use of the product or process—customer sites, call listening, ride-alongs, time-and-motion at the receiver’s workplace.

• Strengths: reveals what people do, not only what they say; catches workarounds
• Weaknesses: Hawthorne effect if intrusive; needs skilled observers; may miss latent needs
• Green Belt use: validate survey claims (“easy to use” vs 12 clicks observed); map internal customer struggle

Triangulation: when survey scores, interview stories, and observation agree, confidence in CTQs rises.

Writing Questions: Kill Vagueness, Ambiguity, and Bias

Garbage in, garbage out applies to VOC as much as to gage R&R. Train yourself to spot bad items.

Vagueness

Vague questions invite inconsistent interpretation.

Define who, what, when, and the scale. Operationally define terms like “on time” (by confirmed ship date? by customer dock date?).

Ambiguity and double-barreled items

Double-barreled questions ask two things at once:

• Bad: “Was the product affordable and reliable?”
• Fix: Split into separate affordability and reliability items.

Ambiguous pronouns and time windows also fail: “Did they fix it quickly?” — who is they, what is it, what is quickly?

Bias: leading, loaded, and socially desirable items

Also watch order bias (early questions prime later ones) and sampling bias (only happy email subscribers surveyed).

Worked rewrite

Original survey item: “Don’t you think our new portal is better and easier than before?”

Problems: leading (“Don’t you think”), double-barreled (better and easier), vague (“better”), biased toward yes.

Improved pair:

1. “Compared with the previous portal, completing a routine address change in the new portal is: Much harder / Somewhat harder / About the same / Somewhat easier / Much easier / Not applicable — I have not used both.”
2. Optional open: “What, if anything, made the address-change task harder or easier?”

Designing a Minimal VOC Collection Plan

1. Purpose: What decision will the data support (prioritize CTQs, validate charter Y, check improve)?
2. Customers: Segments from II.B.1 (internal vs external, primary vs secondary).
3. Methods: e.g., 12 internal handoff interviews + observation of 10 transactions + short external survey n≥100 if breadth needed.
4. Instruments: Scripted interview guide; survey with defined scales; observation checklist.
5. Sampling: Who is in/out; time period; how you avoid only angry or only loyal voices.
6. Bias control: Pilot the survey on 5 people; revise ambiguous items; train interviewers; record coding rules for open text.
7. Analysis path: Affinity themes → frequency/severity → candidates for CTQ trees (II.B.3).

Exam Application Patterns

• Choose observation when stated satisfaction conflicts with field failures or when customers cannot articulate workarounds.
• Choose interviews for few strategic B2B accounts or complex internal customers.
• Choose surveys when you need comparable scores across segments.
• Choose focus groups to explore language before a large survey.
• Reject options that use leading or double-barreled questions as “good VOC design.”
• Prefer specific, measurable phrasing that later becomes a CTQ measure.

Open-book tip: You rarely need to memorize survey software. You do need to diagnose bad questions and match method to purpose. That is the Apply skill ASQ is testing in II.B.2.

Collect VOC like a measurement system: clear definitions, controlled method, and minimal bias—so Define-phase requirements rest on evidence, not anecdotes from the loudest meeting.