I worked on a very popular free desktop product (BitTorrent/uTorrent) when a major revenue stream was the bundling of crapware in the install path, depending on people clicking the dialogs without reading them. I remember an executive announcing in a meeting that "search engine replacement" (reconfiguring your browser and/or system to a shadier alternative that made us money) was viable because the telemetry data showed that not very many people uninstalled in the 2-3 weeks after this was done to them, so, yay! another revenue stream people don't seem to totally hate! But they had read the tea leaves wrong and the backlash eventually caught up with them. It shows that reading mountains of data (100 million+ active monthly users) and reaching solid conclusions about user sentiment is a black art and numbers can be presented to show whatever point of view you want.
There is a sizable middle ground between "basic intuition" and "black art".
I think anyone with some experience in that type of software would intuitively understand the negative user experience described here.
Seems like a pretty straightforward case of the classic "Unless their salary depends on not understanding" rather than some opaque wall of unknowable unpredictable consequences.
Yeah, the sizable middle ground is I believe what props up most of the adtech industry. The sizable middle ground is occupied by people with minimum understanding of statistics, who bullshit themselves and each other with data - but as long as nobody can obviously tell they're wasting money, they're all happy and the money keeps flowing.
