What's the best way to present NPS survey questions to users on phones?

Question

NPS surveys generally take the format of:

The touch area of each selection seems to be pretty tight when I put my mockup on a phone... Is there a better design pattern that I could be using here for this type of input?

Other options I considered:

SLIDER

standard slider: http://opinionmeter.com.pk/wp-content/themes/om/images/TouchPointMobileScreeshots/iPadNPS-Large.jpg

fun slider: https://dribbble.com/shots/1641720-NPS-Survey-Interface/attachments/257028

Interaction wise, sliders could work because they can snap to discrete values. The sliding action makes it easier to fine-tune values than tapping since the user can adjust their input provided the slider gives appropriate feedback.

I didn't like this option because sliders tend to have a default starting value and I felt that this would bias the survey and many users may just submit the default value in order to bypass the survey instead of actually making a selection.

PICKER

http://cloudlink.soasta.com/soasta/attachments/soasta/TT-TestExecutionAndManagement/390/1/UIViewPicker.png

Pickers tend to have an empty or default field that the user must tap in order to bring up the picker UI. I couldn't think of something reasonable to put in that field... I think the interaction is much more straightforward with the picker than with a slider though...

Answer 1

1-to-10 entry can be accommodated on phone screen as long as you use zoom-in key interaction design to facilitate entry.

For NPS, I think it's important to display all 11 numbers in left-to-right sequence, and provide d a one-click way to pinpoint the exact number, otherwise you may inadvertently introduce biases towards certain range.

Below is a mock-up of 3-states: thumb-hover, thumb-down, and thumb-up. It's important to make sure during thumb-down, user can slide-left and right to select adjacent values.

(sorry, I forgot to add 0)

Answer 2

Mobile phones do not have a huge screen, even phablets! I think it is better to use a 5-point scale or likert from unlikely to very likely for instance. A10-point scale, as i see in the first pic, would be more complicated for users to complete. I use 7 or 5-point scales in ux artifacts like nps for many reasons, being odd scales better than even for analysis purposes.

Answer 3

Keep people at the center of data collection. People easily relate to a 10 point scale. For example, all TV reality shows have a rating scale of 10. People subconsciously relate to brands using a 10 point scale. Although the conversation is open for debate, I have seen 10 point scales work best for research studies I have done in past.

Below are some screen shots on how you can present a 10 point scale on different devices.