Without any further information from that happiness survey, it looks like a fairly unclever feature to me.
If I task-switched in after frustration from using another application I might as well enter No. Or if I just had a great coffee maybe I'll enter Yes.
If it is really clever, please show some more stats. 90% said they were happy. Was this correlated with time of day? Or a particular day? More interesting would be something like, on average, how long did it take for users to respond to the survey after it was activated. I venture to guess that this information -- unless most users give detailed comments, as opposed to clicking Yes/No, which most lazy users would do -- is hard to make useful unless you collect further usage details which would put the app into the spyware category.
The question should probably say, "are you happy with Xobni?"
That said, the feedback from this kind of thing can be surprisingly useful. You get a feel for the general mood of your users, and learn what the common pain points are (unhappy people generally say why).
Looking clever and being clever aren't always the same...
Yeah, and it only takes one feedback from one user to make you realize there's a real problem with something you've recently changed, which is actually what you want--if you can fix something before most users see it, there was no problem!
I will agree with the grandparent that the most obvious part of that feature isn't particularly clever: having a feedback field is a good idea that has been done before.
Also, the question may or may not be a bit vague. (I kind of like it because it gives a pulse on their software, especially on an individual user who could be sent an e-mail from a help desk technician when suddenly they choose no after selecting yes for many months.)
Either way, asking users if they're happy is a good idea because they will have statistics to prove it. For example, when I helped users with their support questions, I automatically asked them to fill out a survey. The question "would you recommend this to a friend" yielded a 9 out of 10 agreement rate (93% I think) over a period of a year. I am proud and have that as the first thing on the front page on my web site.
I think web apps + good stats are the way to go, so I'm gearing for that rather than desktop apps.
IMO, the most clever thing about this is that the team at Xobni is admitting (and building into their product development) the fact that users are oftentimes a helluva lot smart about what users want than they are (imagine that!).
Too many startups give lip service to the "release early, release often" mantra, but fail to do it... It's a lot safer to keep your hypotheses behind the curtain, I guess.
The other clever thing that they did was a good balance of making the feedback mechanism lightweight but obvious. We've done this with RescueTime and have literally gotten thousands of feedback notes.
Survey questions like this one are not effective, monitor the usage & behavior of your users, which features are they using , which are they ignoring.
Compare the usage of users which are happy with your product (defined as continue using your software for a long time) with users that dump your product after a few hours of playing with it. What features are the happy users using? What are the unhappy ones missing?
Data mine everything, knowing your users behavior is power.
Wow, I think I want to try that. Not getting feedback from users is a major frustrating point. I could imagine some would be prepared to at least press the Yes or No button. And once they made that step mentally, perhaps they'll even leave a comment.
If I task-switched in after frustration from using another application I might as well enter No. Or if I just had a great coffee maybe I'll enter Yes.
If it is really clever, please show some more stats. 90% said they were happy. Was this correlated with time of day? Or a particular day? More interesting would be something like, on average, how long did it take for users to respond to the survey after it was activated. I venture to guess that this information -- unless most users give detailed comments, as opposed to clicking Yes/No, which most lazy users would do -- is hard to make useful unless you collect further usage details which would put the app into the spyware category.