Objectivity, Subjectivity and Metrics in Software Products
While developing software, one of the central points that take an enormous amount of time and effort is determining whether the changes you're deploying actually do improve the product.
You can feel and have the opinion that what you've done is an improvement. Still, it's important to recognise your own vested interest in painting your own effort in a good light renders your own opinions biased and unreliable.
The tool for solving that conundrum is analysing option A and option B objectively through some dimension. For example: “how often did end-users engaged while in version A versus version B?” or “how many did actually come back the next day or week?”.
These quantities can be measured objectively and presented to persuade anyone who might take the opposite position (on an opinion basis).
Note that after a set of objective metrics is measured, the end decision is often still subjective. Data only informs, it does not decide what's important.
If a business goal is to retain users, taking a short-term engagement hit to deliver a more premium and trustworthy experience can be a strategy. On the other hand, if the business goal is immediate extraction, more annoying popups and nagging will motivate short-term engagement at the cost of driving away and alienating another set of users.
These are decisions and trade-offs.