Kano model: distinguish customer expectations from attractive extras

The Kano model distinguishes requirements by how their fulfilment and non-fulfilment relate to customer satisfaction. A basic expectation can create dissatisfaction when absent without producing much extra appreciation when fulfilled. A performance requirement can matter in both directions. An attractive requirement can be welcomed when present without being expected when absent.

Those categories are hypotheses to investigate with customers. A workshop in which colleagues label their favourite feature a “delighter” has not carried out Kano research. The model also does not provide a complete priority order: cost, importance, reliability, accessibility and the wider service promise still matter.

This guide follows a hypothetical independent hotel considering an early-departure breakfast pack. Its questions, counts and decisions are invented teaching examples. The guide shows how to examine a specific requirement while keeping disagreement and uncertainty visible.

Where the model comes from

Noriaki Kano and colleagues’ 1984 paper, “Attractive Quality and Must-Be Quality”, proposed a two-dimensional view of quality and distinguished attractive, must-be and one-dimensional elements. The publisher’s English abstract describes questionnaire applications to televisions and clocks. It does not establish that every modern service feature fits one stable category.

The familiar diagram separates the degree of fulfilment from satisfaction. Its curves are conceptual illustrations rather than measured response functions for your product. The Double Diamond can place this inquiry after the team has developed a credible understanding of the problem.

Start by understanding customer needs. The hotel might discover that some guests leave before breakfast service begins. That circumstance suggests a requirement to examine; it does not yet prove a preference for a particular bag of food or a new room package.

Understand the categories without assigning them by intuition

The three familiar categories describe different response patterns. Keep the additional outcomes visible when coding a study; they are not inconvenient answers to remove.

Category	Meaning within the model	Important caution
Must-be, or basic	Absence dissatisfies; fulfilment is expected	Does not mean every customer expects it
One-dimensional, or performance	Better fulfilment pleases; poor fulfilment dissatisfies	Not a calibrated linear effect size
Attractive	Fulfilment pleases; absence is tolerated	Does not establish importance or willingness to pay
Indifferent	Presence and absence provoke little preference	Check whether the requirement was understood
Reverse	The proposed presence is less welcome than its absence	A genuine preference can differ from the team’s assumption
Questionable	The answer combination warrants checking	Do not silently recode it as indifference

Berger and colleagues’ 1993 Kano compendium explains the categories, paired questions and evaluation table. It also discusses alternative answer formulations and coding choices. Preserve the exact questionnaire and coding version you use rather than combining versions during analysis.

For the hotel, “a clean room” may sound like an obvious basic expectation, but the teaching label is not a finding from this imagined study. Required operating standards should be met regardless of whether a satisfaction questionnaire labels them must-be. A customer-preference model is not permission to compromise essential obligations.

Define one requirement in a real use context

Write what is provided, for whom, when and under which conditions. “Excellent breakfast flexibility” is too vague. “A pre-ordered breakfast pack available at reception at 06:00 for guests departing before the dining room opens” gives people something recognisable to assess.

Avoid combining several changes in one item. An early pack, free room service and a lower price cannot be classified as one meaningful requirement. A respondent could like one part and dislike another. Likewise, words such as delicious, convenient and premium can lead the answer before the person considers the offer.

Use customer interviews to understand departure routines and current alternatives before presenting the feature. Some guests may buy food at the station, some may not want breakfast and some may have needs the proposed pack does not serve. These are relevant conditions, not resistance to the concept.

A concept brief should state the price arrangement, ordering deadline and collection process clearly enough for the question to be interpretable. The hotel must avoid promising flexibility that reception cannot actually deliver.

Ask about presence and absence of one requirement

The functional question describes the requirement being provided. The dysfunctional question describes its absence. Both concern the same requirement and circumstances, not a good version versus an unrelated bad version.

For our invented hotel, a draft pair is: “For an early departure at this hotel, how would you feel if a pre-ordered breakfast pack were available at reception at 06:00?” Then: “For the same departure, how would you feel if that breakfast pack were not available?” Keep the stated price arrangement identical in the surrounding scenario.

Kano’s familiar response options distinguish liking the situation, expecting it, neutrality, being able to tolerate it and disliking it. These are semantic categories, not numbers on a satisfaction scale. Use a consistent, pretested wording and the corresponding evaluation table. The descriptions here explain their meaning rather than supplying a validated translated questionnaire.

The Government Analysis Function’s questionnaire guidance supports clear, balanced questions and cognitive and usability testing. Pretest whether guests understand both scenarios and distinguish an expectation from a preference. Do not help one guest answer by calling the pack a bonus while describing it to another as a standard entitlement.

Choose a sampling approach that fits the intended conclusion. A small exploratory group can examine comprehension and variation. It cannot establish the distribution of categories across all future hotel guests merely because every person answered both questions.

Code each pair before summarising the requirement

Combine the two answers for each respondent using the selected evaluation table. Do not classify the requirement using the answer to the presence question alone. Someone liking a pack could regard its absence as acceptable or unacceptable; those combinations have different meanings.

The following table shows selected combinations in the conventional coding arrangement described by the compendium. The answer descriptions are paraphrases for explanation. It is not the complete matrix or a substitute for retaining the original instrument.

Reaction to presence	Reaction to absence	Conventional code
Welcomes it	Feels neutral	Attractive, A
Welcomes it	Dislikes it	One-dimensional, O
Expects it	Dislikes it	Must-be, M
Feels neutral	Feels neutral	Indifferent, I
Dislikes it	Welcomes it	Reverse, R
Welcomes it	Welcomes it	Questionable, Q

Retain incomplete pairs and questionable outcomes as such. Check whether an apparent contradiction reflects misunderstanding, an unsuitable description or a recording error. A reverse preference can be substantive: a guest might prefer a lower-complexity service without a pack or object to the arrangement implied by it.

Build the distribution for each requirement. Preserve the counts and recruitment context before choosing any overall label. The modal category is the most frequent category; it need not represent a majority or a clear consensus.

Read disagreement before choosing a category

Suppose an invented exercise produces 40 complete pairs: 14 attractive, 10 performance, 10 must-be, four indifferent, one reverse and one questionable. These sum to 40. Attractive is the largest category, but 14/40 is only 35%. Calling the breakfast pack “a delighter for our guests” loses most of the variation.

Kurt Matzler and Hans Hinterhuber’s 1998 paper explains frequency-based classification, more differentiated interpretation and satisfaction coefficients. It describes how expectations can differ between customer segments. Its ski-industry application is not evidence that this hypothetical hotel’s categories are correct.

Investigate useful differences tied to the task. Guests leaving before breakfast may have different expectations from those with a relaxed morning. A guest who did not choose a room rate including breakfast may interpret the offer differently again. Do not invent segments after the fact merely to make the desired feature look attractive.

Report category distributions within meaningful groups when the design and numbers justify that analysis. For small exploratory groups, preserve the accounts and treat the difference as a question to test. A one-person subgroup cannot settle a market segment’s requirements.

Expectations can change as the service and competitive context change. Revisit the evidence when those conditions change. Do not treat a common story about delighters eventually becoming basics as a guaranteed lifecycle for every requirement.

Calculate coefficients without predicting a satisfaction increase

The conventional satisfaction coefficient is (A + O) / (A + O + M + I). The dissatisfaction coefficient is −(O + M) / (A + O + M + I). Reverse and questionable responses are excluded from these denominators but should remain visible in the report.

Using the invented counts, the denominator is 14 + 10 + 10 + 4 = 38. The satisfaction coefficient is 24/38, approximately 0.63. The dissatisfaction coefficient is −20/38, approximately −0.53. These are summaries of coded answers under that calculation, not a prediction that implementing the pack raises satisfaction by 63%.

Explain exclusions, missing pairs and the coding version. If the denominator is zero, the coefficients are undefined rather than zero. If reverse or questionable outcomes are common, that is a reason to investigate the question and preferences instead of presenting neat coefficients that hide them.

Use the calculation only if it helps the decision. It can show that an aggregate label conceals both appreciation and dissatisfaction potential. It does not transform an inadequate sample or a leading questionnaire into dependable evidence.

Combine category evidence with importance and delivery

Mikulić and Prebežac’s critical review examines classification methods and warns that a Kano category does not itself establish a requirement’s importance in overall satisfaction. It also discusses how operationalising the question can change what is classified.

For the hotel, compare the breakfast question with reliable room access, accurate departure information and the current offer’s obligations. An attractive pack could be a minor preference while an unambiguous check-out arrangement matters much more. Idea evaluation should keep benefit, evidence and constraints separate.

Examine delivery assumptions through assumption mapping. Can reception identify the order, respect the agreed timing and handle an unavailable item? A favourable category distribution cannot answer those operational questions.

A concept test can investigate the ordering and collection experience. A separate limited rehearsal can examine operations. Connect those next steps to staged funding rather than treating Kano as approval to launch the whole service.

Evidence and limitations

Löfgren and Witell’s 2008 review examined 33 papers and raised methodological concerns about some developments in the theory. The accessible journal summary supports that historical caution; it does not describe the complete current research literature.

Kano provides a useful distinction between appreciation and expectation. Its results remain sensitive to requirement definition, questionnaire interpretation, recruitment, segment composition and coding choices. It does not prove purchase, retention, profit or a causal improvement in overall satisfaction.

The hotel example is fictional, including every count and coefficient. Use the model when a clear requirement deserves investigation through paired scenarios. Finish with a category distribution, its evidence limits and a proportionate next decision, rather than a colourful label that substitutes for customer research.