I am working on a user feedback mechanism, and have seen conflicting content out there on the best way to remove user-friction from submitting useful feedback.

There is the simple and streamlined process (like Kampyle or Uservoice) where all of the feedback form sits in one place and users only have one grouping to look at.

Then there is the Microsoft send-a-smile model, where users choose an emotion before being offered a text box to submit more information. While this at first seemed counter-intuitive to me, there seems to be some user studies and research papers supporting it (section 2.2 of this paper).

What else is out there on this one-phase versus two-phase feedback mechanism experience? In which situations is one better than the other?

1 Answer 1


The goal

Let's make the goal obvious: we need to get feedback from users, so the success metric is conversion rate, which defines the best option among the ones.

Some heuristics

  1. People don't like to fill the forms, so minimize input
  2. People don't like complex tasks, so make it as simple, as possible
  3. People estimate complexity of the task by assessing UI. This is possible place for applying progressive disclosure UI pattern, or split the form
  4. Complex task is percieved more easy if it's divided by chunks.
  5. To engage users, place more easy task first, then complex one. So first screen is good for rating (easy interaction), and second one is for text input.

And use A/B testing to evaluate results.


Example 1
enter image description here

  • This form looks complex as there are too many input fields
  • User need to decide on the feedback subject. Don't make them think, it's your task to classify the feedback!

Example 2
enter image description here

  • The form is structured, so you can jump to the needed section (Audio or Video)
  • Input is minimized using checkboxes
  • Text input box is a bit unclear, though

Example 3
enter image description here

  • Input is minimized using radios
  • Optional text input

Example 4
enter image description here
Very nice two-step solution

Example 5
enter image description here

  • First step allows to focus on text entry with very clear UI
  • Step two is for additional info
  • Also note that in a number of cases you can receive value even if user only answers first part of question. e.g. (1) How do rate our service? (2) Why did you give that rating? Even if just record at step (1) that is still useful data. Thus by the heuristics given in answer a two-step is beneficial in this case. In a case where it's all or nothing response, then if user feels they were "suckered-in" to a long process they may abandon a two step leaving you with no data.
    – Jason A.
    Jan 24, 2016 at 23:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.