Objectives and Key Results. The things you choose to optimize for during any given planning period.
OKR: Improve customer feedback ratings from 4.2 to 4.5+ stars
Result: Customer support personnel start handing out free upgrades/bonuses/merchandise to complaining customers, resulting in many 5 stars reviews and while not solving any problems customers are facing, still achieving the OKR.
Measuring performance by numbers instead of personal judgement (from managers/team leads) is problematic.
Edit: It certainly makes sense in some departments (like decreasing page load time from 2 seconds to 1[0] but forcing that across all verticals is retarded.
[0] then again you could achieve that for example by deleting all JS from the site
/rant
In your example, the "O" would be something like "provide a good experience for app users" and a "KR" would be "improve star ratings from 4.2 to 4.5". The "O" helps keep you focused on what you actually care about, making the metric subservient to that goal.
Improved feedback rating is a result, not an objective on its own. It could be filed under, say, "improve customer experience with the app". That objective would have many (3-5) results that you track, with improving feedback being just one of them. Others could be improving page load times (as you mentioned), adding critical missing functionality, or whatever else. These results can also be based on subjective opinions, but need to be somewhat measurable (otherwise how do you know whether you made things better?) Ultimately if you made good enough progress in each of these KRs, you can consider the objective achieved.
Whatever alternative methods (judgement of managers) is up for debate though, I didn't say I had a good one. Just that OKR's in my mind are a waste of time for the most part, especially when you make the whole organization do them.
And look at products like Facebook, they seem to be built solely on OKR's.