The Kano Model: Prioritize Features by How Users Actually Feel
By Road map ROI, from Cookie Tech, which builds custom software · Published Oct 3, 2026 · Last updated Nov 3, 2026
What changed: Rewrote the closing section to match what our calculators do.
Five categories. Two survey questions. One classification per feature, and a clear verdict on whether to ship, invest, or kill it.
Noriaki Kano and colleagues published the model in 1984 (Kano et al., 1984). Its starting point: not all features matter the same way. Some are invisible until they break. Others scale with effort. A few can delight. The Kano Model gives you a repeatable way to tell them apart, using a two-question survey, a 5×5 matrix, and a verdict you can take to your roadmap.
Interactive
Kano Classifier
Pick one feature. Imagine a representative user. Answer both questions and the model classifies the feature for you.
Classification
One-dimensional (Performance)
Invest. Satisfaction scales linearly with how well you execute it.
The five categories
- Must-be (Basic)
- Table stakes. Their absence creates dissatisfaction. Their presence doesn't move the needle. Login that works, search that returns results, an invoice that arrives.
- Performance (One-dimensional)
- Linear satisfiers. The better you do them, the happier users tend to be. Speed, battery life, accuracy, price.
- Excitement (Attractive)
- Unspoken delighters. Users don't expect them and won't complain when missing, but their presence can delight. A possible way to stand out.
- Indifferent
- Features users genuinely do not care about. Building them spends effort without changing satisfaction.
- Reverse
- Anti-features. Their presence actively reduces satisfaction (think aggressive notifications, forced AI features, dark patterns). Remove on sight.
The two-question survey
For each candidate feature, ask the same user two questions: a functional form ("If this feature were present, how would you feel?") and a dysfunctional form ("If this feature were missing, how would you feel?"). Use the standard five-point scale: I like it · I expect it · I'm neutral · I can tolerate it · I dislike it. A small sample, such as 20 to 30 users, is our suggestion, not a rule. The classifier above shows the resulting category for one pair of answers.
When Kano beats RICE and FVI
RICE and FVI assume every initiative on the backlog is a legitimate candidate; they rank what's already in the list. Kano is upstream: it tells you whether a feature belongs in the list at all. Use Kano early, when the question is "is this even worth scoring?" Use RICE for weekly grooming of the survivors. Use FVI when finance needs the dollar number behind the rank. The three frameworks compose; they don't compete.
Common failure modes
- Surveying the wrong users. Results from power users may overclassify Excitement features. Sample the segment you actually serve.
- Forgetting the decay curve. Kano described how today's Excitement can become tomorrow's Performance and eventually a Must-be (Kano, 2001, as summarized on the Wikipedia page above). Consider re-running Kano on long-lived features, for example once a year.
- Building too many Excitements. A roadmap of nothing but delighters can mean the basics are quietly breaking. Watch the balance with Must-be and Performance work.
Sources: Kano model, Wikipedia (Kano et al., 1984; Kano, 2001). The classification table is the common Kano evaluation table; we have not matched it cell by cell to a published source. Sample size and the re-run advice are our suggestions.
Frequently asked questions
- What are the five Kano categories?
- Must-be (Basic), Performance (One-dimensional), Excitement (Attractive), Indifferent and Reverse. Must-be features cause dissatisfaction when missing, Performance features raise satisfaction the better you do them, and Excitement features delight users who did not expect them. Indifferent features are ones users do not care about, and Reverse features reduce satisfaction when present.
- How do you run a Kano survey?
- For each feature, ask the same user a functional question ('If this feature were present, how would you feel?') and a dysfunctional question ('If this feature were missing, how would you feel?'), each on the five-point scale from 'I like it' to 'I dislike it'. Read the category from the 5×5 matrix. A small sample, such as 20 to 30 users, is our suggestion, not a rule.
- When should I use Kano instead of RICE or FVI?
- Use Kano early, when the question is whether a feature belongs on the backlog at all. RICE and FVI rank what is already on the list, so use RICE for weekly grooming of the features that survive and FVI when finance needs the dollar number behind the rank.