Prioritising within categories

What about features that are in the same category?

When the Kano categories are clear-cut, prioritising is easy. If 9 out of 10 answers categorise feature A as a Natural and feature B is an Indifferent 9 out of 10 times, it's easy to make a decision about what features to work on first.

But what about features that fall in the same category?

Suppose you only have time to work on one of three features. And suppose these are the results of your survey:

Feature	N	O	A	I	Category	Priority
A	53	24	2	21	Natural	?
B	46	20	27	7	Natural	?
C	38	28	18	16	Natural	?

In the classic approach, you'd take on feature A first, because it has the most Natural answers. But feature A also has the most Indifferent answers. Is that really worth deprioritizing the other features for? (Actually, this kind of result also means you should look for segments in your customer base, but let's start with prioritising).

Berger (1993) describes a way of ranking features based on their total impact on customer satisfaction. It's called the M>O>A>I rule. (I'll use Must-Be here instead of Natural, because that's the category label the researchers used).

You apply this ranking by listing categories by frequency:

Feature	Most frequent	Second most frequent	Third most frequent	Fourth most frequent	Priority
A	M	O	I	A	?
B	M	A	O	I	?
C	M	O	A	I	?

Then you rearrange the rows into groups according to the M>O>A>I order.

So the feature where Must-Be is the most frequent category gets listed first. But when two or more features tie, you look at the second most frequent category. If the second most frequent category for the first feature is One-Dimensional and for the second feature Atrractive, the first feature gets ranked higher than the second.

Using this rule, priorities are assigned thusly:

Feature	Most frequent	Second most frequent	Third most frequent	Fourth most frequent	Priority
A	M	O	I	A	2
B	M	A	O	I	3
C	M	O	A	I	1

C is the feature you should start working on first in this case. It has a less "Must-Be" answers than A, but its total impact on customer satisfaction is higher. The total of Must-Be, One-Dimensional and Attractive response categories is higher for feature C than it is for feature A.

Of course, if none of the tied features are in the Must-Be category, you apply the same rule, leaving out the category they're tied in.

In this example, all features are Attractive:

Feature	M	O	A	I	Priority
Feature A	2	24	53	21	?
Feature B	27	20	46	7	?
Feature C	18	28	38	16	?

Ranking means simply applying the M>O>I rule (because A is already the most frequent):

Feature	Most frequent	Second most frequent	Third most frequent	Fourth most frequent	Priority
A	A	O	I	M	3
B	A	M	O	I	1
C	A	O	M	I	2

Were you to use the classic approach and just count the Attract answers, feature A would have been at the top of your list. But because it has so few Must-Be answers, feature A is now at the bottom. Features B and C have more overall impact on customer satisfaction, so they must be prioritised.