How to ask a one-dimensional / performance attribute question in a Kano survey?

Question

I am putting together a Kano analysis survey, and want to understand how to phrase the question for a performance attribute.

In putting together the survey, for each attribute I would ask the two standard questions:

1. How would you feel if this feature is present
2. How would you feel if this feature is not present

Each question would have five possible answers, from "I like it" to "I dislike it".

My question is this - how should I phrase these survey questions for a performance feature - where the implicit assumption is that the 'feature' is always present, but may be poorly or well implemented.

As an example, let's take the fuel consumption of a car. Fuel consumption is neither present nor absent. Should I instead ask "How would you feel if the car had poor fuel consumption" and "How would you feel if the car had best-in-class fuel consumption"?

Answer 1

First, you shouldn't assume a particular feature or attribute of a product will end up being classified as a Performer feature. It could be that enough time has passed and enough standardisation has occurred that the requirements for that particular feature has stabilised at some plateau level, and the customer satisfaction response now follows the Basic Expectation curve.

For example, there was a time when the number of pixels a digital camera captures was a Performer feature - the more the better, because there were many cheap digital cameras that were puny in their specs. Today though, for consumers, this attribute is closer to being a basic expectation - takes enough that you can't see the grain, and not so much it takes ages to download one image. Whatever that means. Sure, there are digital cameras with best in class pixel resolutions, but there's still a very big market for "good enough" pixel resolution. Using the Kano Model will help establish if your "good enough" camera can be improved and enhanced into a Extreme Plus Pro™ version, or whether you should focus your attention onto other factors.

If you believe a given feature will result in a classification of being a Performer feature, then you should be expecting the function form of the question to result in "I like it that way" (and not "I expect it that way" or lesser alternatives), and the dysfunctional form to have "I dislike it that way" (and not "I can tolerate it that way" etc responses).

One possible problem with your suggested phrasing is that you've expressed both the functional and dysfunctional forms as specific levels: "poor fuel consumption" and "best-in-class fuel consumption". The difficulty here is that these phrasings may not be equally weighted opposites in the customers mind.

Given all that, your best course of action might be to phrase the questions as follows:

1. How would you feel if [the feature] was good enough?
2. How would you feel if [the feature] was not good enough?

Note how (a) the description falls into a vague middle ground, and (b) dysfunctional form is literally the negated functional form (i.e. opposite).

With (a), you won't have a big problem with misleading false-negatives where a customer doesn't choose "I like it that way" because they happen to be thinking "Oh, if it's the bee's knees it might be too expensive" (i.e. they're second-guessing knock-on effects onto other attributes).

With (b), you side-step any complications arising from customers who think the opposite of "best-in-class [attribute]" is "'effing woeful [attribute]", with your phrasing of "poor [attribute]" being interpreted more softly as "kinda sub-par". If they read "poor" in that manner, they're more likely to answer "I tolerate it that way" instead of "I dislike it that way" (which you should be expecting if the feature is in fact the Performer class of features).

Answer 2

Also try an alternative approach. Feeling like/dislike spectrum are typically "Objection boring & leading the witness" during surveys anyway. Plus people often project their responses to surveys in a way that doesn't always match up to their actual behavior.

In the question above you're really trying to get to what I can only assume is feature importance or to determine if the user finds value in the feature.

Another way to ask for the same result is to try a $100 spend survey. Basically you feed in a mixed bag of features that you find are a little close to the one you're asking and ones that are low on the users needs and assumed "high" aswell. You tell them they have $100 to spend and how much money would they spend on improving each feature in the list.

As they spend their $100 they are really "voting" on the level of importance per feature and from this you could infer their base "feeling".

If you want to also then drill down on the health index of that said feature you can then ask another set of questions using the same approach by using humour and/or injecting emotion into the equation .. "You have 10 likes and 5 punches, below is a list of 15 features"

Now what this would do or a formula like this it essentially forces the user into having both positive & negative emotions towards the feature(s) basically you're saying outloud "I am going to kill 5 of these things you may love but keep 10...so... who lives who dies.." why force a negative? people are typically optimistic or pesimistic in nature during survey's anyway so either way you end up with a set of data that you really aren't sure at times if they are being "nice" or being truthful. Forcing them to have both sets of emotion(s) tends to draw out their raw choices.

Its what i call "kicking the sick puppy" survey(s)... but i find them effective :)