The reasons are not obvious since we directly estimate the number of good and bad events that have affected users without considering emotional weight of those events. One bad event could easily lead users to a negative opinion about the product even if it happened after a dozen of good ones.

Users evaluate the product at different time ranges than we do. We may have already fixed the bugs, but the users who shared feedback were thinking of their negative experiences in past.

We made a series of minor mistakes in a short amount of time. Due to the frequency (not the seriousness) of errors, users' opinions about the product temporarily dropped. If the errors did not have serious consequences for the users, then according to they will soon forget about it, and the product will restore its reputation.

Our users have recently used the products of our competitors. A brighter wrapper, or a technological perfection of some details they saw there, distorted their opinion about our product.

Our application caused intense discomfort to the user. This triggered a powerful emotional response that was remembered as the "peak" of the entire product experience. For example, the application accidentally opened a pop-up with the question, "Are you sure you want to delete your profile and all data permanently?". The user, of course, clicked "No," but the emotion he experienced at that moment left a significant imprint on his evaluation of the product. According to , this memory will fade over time, but we need to make sure that such errors do not occur again.

. Perhaps, with their complaints, our users mask the fact that it is difficult for them to work with the product (low qualification/age/the habit of working with paper documents).

. The user started using our product only to confirm some of his beliefs. His complaint is not useful feedback but rather a form of goodbye to us.

. The product components we enhanced were not in the scope of our users' main interests.

. The problem is not in the product quality but in its complexity. Perhaps the product is of high quality in terms of the tasks it solves, but at the same time, it isn't easy to use.

If we see a steady increase in frustration but still do not understand the reason, it may be helpful to temporarily provide users with some kind of "Report a problem" button ().