What needs to be finished first
A clearly formulated list of 10-25 feature candidates or improvement ideas is available, each described as an unambiguous sentence.
A preliminary understanding of the target audience and their expectations is available, so features can be evaluated in the right context.
What needs to be ready before start
Online survey tool (Typeform, Google Forms, SurveyMonkey) or interview guide; Kano evaluation spreadsheet with discrete analysis table (5x5 matrix) or continuous analysis plot; feature list as a numbered catalog.
One researcher owns the survey and analysis, one product manager owns the feature list and interpretation, one designer supports feature wording, and at least 30 respondents from the target group for statistical validity.
Feature list with short neutral descriptions; target group definition and recruiting plan; existing customer feedback data as validation; possibly a budget for incentives.
1-2 days setup, 1-2 weeks surveying, 1-2 days analysis
Create a paired question set for each feature (functional: "How do you feel if feature X exists?"); dysfunctional: "How do you feel if feature X does not exist?"). Set answer scale to 5 levels (like, expect, neutral, can live with, dislike). Run a pilot with 3 people, then scale up.
The one question this method answers
Which of the proposed features are must-have basics, performance drivers, or delighters, and which are irrelevant or even blocking?
Marker: Phase
| Step | Duration | Action | Hint |
|---|---|---|---|
1Phase 1: Feature list and question pairs | 2-4 h | Formulate each feature description neutrally and concretely. Create both functional and dysfunctional questions. Set answer scale. Randomize order. | Marketing language will distort the result. Prefer technical-neutral phrasing: "inline validation in form" instead of "seamless, intelligent form help." |
2Phase 2: Pilot with 3-5 people | 2-4 h | Test the survey with a small group. Check comprehensibility. If systematic understanding issues appear, rephrase or drop the feature. | Skipping the pilot reduces data quality later. At least 3 participants, ideally 5, with a follow-up interview to review question clarity. |
3Phase 3: Main survey | 1-2 weeks | Distribute online survey to at least 30 target-group people. Track response rate, and recruit reminders or improve incentives where needed. Check sample representativeness. | Fewer than 30 responses make the analysis statistically weak. For a specific subgroup, 15-20 can work, but only with qualitative interpretation. |
4Phase 4: Discrete analysis | 4-6 h | Assign one category per feature via 5x5 matrix (Must-be M, One-dimensional O, Attractive A, Indifferent I, Reverse R, Questionable Q). Determine dominant-category mode per feature. | High Q share usually means the question was unclear or respondents did not understand the feature. If more than 5% are Q for one feature, revise the question. |
5Phase 5: Interpretation and roadmap derivation | 2-4 h | Must-be: required, do not remove. One-dimensional: investment drives satisfaction in a near-linear way. Attractive: differentiation, low expectation. Indifferent: de-prioritize. Reverse: generally remove. Discuss implications with PM and stakeholders. | Kano categories are a snapshot. Delighters become performance over time, then become must-be (for example smartphone camera). Re-run survey every 12-24 months. |
What comes out at the end
Analysis document with feature list, Kano category per feature, confidence indicator (share of dominant category), respondent demographics, interpretation recommendations, and roadmap derivation per feature cluster.
Create one file per survey round with date, sample size, and target-group description. Explicitly document comparison to previous rounds so category shifts are visible.
Compact working template for Kano with context, input, output artifacts, and next step.
| Item | Description | Rating | Evidence | Owner | Next step |
|---|---|---|---|---|---|
| 1 | |||||
| 2 | |||||
| 3 |
What consequence follows from the matrix?
Concrete filled scenario, fictional example
Concrete filled scenario, fictional example
Sample: 67 respondents from online shopper target group (25-55 years, more than 2 online purchases per month). Incentive EUR 15.
| Feature | Category | Confidence | Implication |
|---|---|---|---|
| SSL security indicator | Must-be | 89% | Required, no investment lever |
| Express checkout | One-dimensional | 64% | Performance, higher investment creates higher satisfaction |
| Wishlist | Indifferent | 71% | De-prioritize |
| Live chat in checkout | Attractive | 58% | Delighter, good differentiation |
| Newsletter opt-in mandatory | Reverse | 76% | Harmful, remove |
| Pre-filled demo data | Questionable | 14% | Question wording was unclear, rewrite |
Recognize symptoms and steer against them
Survey runs with 10-15 responses and categories are interpreted as statistically meaningful.
For quantitative claims, use at least 30 responses. With smaller samples, treat results as qualitative signals, not distribution-level claims.
Features like "intuitive UX" or "revolutionary speed" are included in the questionnaire.
Keep feature wording neutral and concrete. Describe the functionality, not the effect. Prefer "1-click checkout" over "blazing-fast checkout."
Respondents are not from the target group (e.g., internal staff or wrong demographics).
Enforce recruitment filters strictly. For specific audiences, use external recruiter support (Respondent, User Interviews) instead of a convenience sample.
A Kano result from two years ago continues to drive decisions although the market has changed.
Re-survey every 12-24 months for core portfolio. Categories move over time (delighter to must-be), and roadmap must move too.
Many features are classified as indifferent but still make it into the roadmap.
Treat indifferent as decision-relevant. Such features offer no satisfaction leverage; challenge whether they should be built at all.
Done signals checkable in under a minute
Go to the profile for purpose, similar methods, and sources or continue to the next method in the catalog.