Segmentation Analysis of the Kano Model

Where a market contains segments with distinct needs, the standard analyses can fail to identify the key are sub-segments of consumers with different views of attributes the Kano Model can give misleading conclusions. Segments can be found using just about any segmentation algorithm. An example is presented.

Preparing data for segmentation analysis of the Kano model

There are many possible ways of preparing data for segmentation analysis of the Kano Model. As with most segmentation studies, there is no correct approach. It comes down to which delivers the most insight into the problem at hand. Two easy approaches are:

• Scale variables: use all the variables created to perform a Continuous Scale Analysis of the Kano Model.
• Difference variables: create variables by subtracting the Functional scores from the Dysfunctional scores, where the data is using its original scale (i.e., values of 1 through 5). The resulting differences will have high scores when Functional scores are greater than Dysfunctional scores and vice versa.

Choice of the segmentation algorithm

Although just about any segmentation algorithm can be used, if using either of the two data preparations described in the previous section, K-Means is appropriate, as its assumptions are consistent with the data preparation.

Where samples are small, a strong preference should be given to choosing a small number of segments (i.e., clusters).

Example

Three clusters were identified using K-Means with the difference variables. The Kano Diagram below shows the Satisfaction Coefficients for the first segment. This group is largely indifferent to all the attributes.

By contrast, the second segment regards all the attributes as being valuable, although there's a nice and clear distinction between which are Must-be, Performance, and Attractive.

The third segment is different again. The biggest difference is that it sits roughly between the first two segments in terms of its overall passion for the category, with all the scores being closer to the Indifferent quadrant.  However, it differs from the second segment in qualitative terms as well. In particular, we can see that Two-day battery life is a clear Must-be in both segments, whereas Workouts is Performance in Segmen 2, and Indifferent in segment 3.