Market research method
Kano model analysis for feature prioritization
The Kano model sorts product features by how they affect satisfaction. Each feature is asked twice — how would you feel if the product had it, and if it did not — and the answer pair classifies it as must-be (expected; absence angers), one-dimensional (more is better), attractive (a delighter), indifferent or reverse.
How results are checked
Named statistical review pending. Selected automated reference cases are published on the validation page.Run a Kano analysis — free, no signup
Open respondent-level functional and dysfunctional answers (codes or text) or paste category counts, and get each feature's category, Fong's test, CS+ and CS− and a satisfaction–dissatisfaction chart. Free and browser-local.
Run a Kano analysis →The paired Kano question
For every feature the questionnaire asks a functional question (“How would you feel if the app shows live courier tracking?”) and a dysfunctional one (“… if it does not?”), each answered on the same five options: I like it, I expect it, I am neutral, I can tolerate it, I dislike it. Many studies add a self-stated importance rating per feature.
The evaluation table below turns each respondent's answer pair into a category. The feature's category is the most frequent one across respondents — provided it is clearly ahead of the runner-up.
The Kano evaluation table
| Present ↓ / absent → | Like | Expect | Neutral | Tolerate | Dislike |
|---|---|---|---|---|---|
| Like | Q | A | A | A | O |
| Expect | R | I | I | I | M |
| Neutral | R | I | I | I | M |
| Tolerate | R | I | I | I | M |
| Dislike | R | R | R | R | Q |
A attractive · O one-dimensional · M must-be · I indifferent · R reverse · Q questionable (contradictory answers).
The method
Formula
CS+ = (A + O) / (A + O + M + I) CS− = −(O + M) / (A + O + M + I) |a − b| ≥ 1.65 · √((a + b)(2n − a − b) / 2n)
where
- A, O, M, I
- the number of respondents classified attractive, one-dimensional, must-be and indifferent
- CS+
- satisfaction coefficient, 0 to 1: how much having the feature raises satisfaction
- CS−
- dissatisfaction coefficient, 0 to −1: how much its absence lowers satisfaction
- Fong's test
- a and b are the two largest category counts out of n classified answers; when the gap is smaller than the critical value, no single category is assigned
Worked example
The synthetic food-delivery app survey that the Kano tool loads as sample data asked 300 respondents about six features (a few skipped a question, so each feature has slightly fewer classified answers):
| Feature | n | A | O | M | I | R | Category | CS+ | CS− |
|---|---|---|---|---|---|---|---|---|---|
| Accurate delivery-time estimate | 295 | 6% | 23% | 48% | 20% | 1% | Must-be | 0.30 | −0.73 |
| Live courier tracking | 294 | 21% | 42% | 16% | 15% | 3% | One-dimensional | 0.67 | −0.62 |
| One-tap reorder | 291 | 29% | 31% | 9% | 27% | 2% | No clear category (One-dimensional lead) | 0.62 | −0.41 |
| Free birthday dessert | 294 | 50% | 11% | 4% | 33% | 1% | Attractive | 0.62 | −0.15 |
| Recipe videos | 292 | 18% | 3% | 4% | 58% | 14% | Indifferent | 0.25 | −0.08 |
| Promotional push notifications | 289 | 7% | 2% | 2% | 54% | 35% | Indifferent | 0.13 | −0.05 |
Synthetic food-delivery app Kano survey · 300 respondents · the same data as “Use sample data”. Category = most frequent answer when Fong's test says it is clearly ahead.
Accurate delivery-time estimate is a must-be: 48% classify it M and its CS− is −0.73, so leaving it out costs a lot of satisfaction while doing it well earns little credit. Live courier tracking is one-dimensional (CS+ 0.67, CS− −0.62): the better it is, the happier users are. Free birthday dessert is the delighter — attractive, with CS+ 0.62 but CS− only −0.15.
One-tap reorder gets no category: 85 respondents say attractive and 89 say one-dimensional, a gap of 4 against a critical value of 18.2. The audience is split, and its CS+ of 0.62 says it is still worth building. Promotional push notifications are indifferent for most, but 35% answer reverse — they actively do not want them — a warning that an average hides.
Using it well
When to use it
Use it when
- Use: Deciding what goes into a release or a product tier: must-bes first, then performance features, then a few delighters.
- Use: Separating features customers take for granted from ones that differentiate.
- Use: Finding features that some customers actively dislike (reverse) before you build them.
Not the right tool when
How to interpret it
Rule of thumb
Build every must-be (high |CS−|) to an acceptable standard, invest in one-dimensional features in proportion to their CS+ and CS−, and add attractive features (high CS+, low |CS−|) to differentiate. A feature with no clear category usually means different segments see it differently — cut the result by segment.
Categories drift over time
Kano categories are not permanent. Yesterday's delighter becomes today's performance feature and tomorrow's must-be as competitors copy it and customers get used to it. Repeat the study when the market moves, and compare categories, not only coefficients.
Questions
Frequently asked questions
What are the five Kano categories?
Must-be (expected; absence causes dissatisfaction), one-dimensional or performance (satisfaction rises with how well it is delivered), attractive (a delighter whose absence is not missed), indifferent (no effect either way) and reverse (some customers prefer not to have it). A sixth result, questionable, flags contradictory answers.
How do you calculate the Kano satisfaction coefficients?
CS+ = (A + O) / (A + O + M + I) and CS− = −(O + M) / (A + O + M + I), using the number of respondents in each category. CS+ runs from 0 to 1 and CS− from 0 to −1; reverse and questionable answers are left out.
What is Fong's test in Kano analysis?
Fong's test checks whether the most frequent category is clearly ahead of the runner-up. With a and b the two largest counts out of n answers, the difference is significant when |a − b| ≥ 1.65 · √((a + b)(2n − a − b) / 2n). If it is not, the feature is reported without a single category.
How many respondents does a Kano survey need?
Enough for each segment you want to report. Categories are proportions, so a few hundred respondents give stable results overall; small segments will often fail Fong's test.
What does 'questionable' mean in Kano?
A questionable answer is contradictory — for example liking both having and not having the feature. A high share of questionable answers usually means the question was misunderstood.
Sources and next steps
References & further reading
- [1]Kano, N., Seraku, N., Takahashi, F., & Tsuji, S. (1984). Attractive quality and must-be quality. Journal of the Japanese Society for Quality Control, 14(2), 39–48.
- [2]Berger, C., Blauth, R., Boger, D., et al. (1993). Kano's methods for understanding customer-defined quality. Center for Quality of Management Journal, 2(4), 3–36.
Run a Kano analysis — free, no signup
Open respondent-level functional and dysfunctional answers (codes or text) or paste category counts, and get each feature's category, Fong's test, CS+ and CS− and a satisfaction–dissatisfaction chart. Free and browser-local.
Run a Kano analysis →