I feel like there are so many factors here that it's hard to identify which thing have the greatest impact. Instead of attempting a coherent argument, I'll offer some further observations.
Taylor Swift is one of the most famous people in the world, yet I know quite a large number of people who could name only one or two of her songs. I would count myself a Taylor Swift fan even though I am in the group of knowing very little of her music. I admire her creativity, business acumen, legs, assertiveness, intelligence, and determination.
In the past, a performer at that height would dominate a much smaller range of media coverage leading to a more profound cultural impact. While being on fewer channels, they'd be on a greater proportion of the whole media landscape.
I think that pushes the dial in both directions. When something is targeted at all, they have to stay around the median to encompass the largest population.
Transformational change happens to a society when something that is targeted beyond the median becomes popular and drags the world with it.
You hear a lot of talk about the Overton window these days. I have heard it raised frequently as an argument for deplatforming. It strikes me as a profound misunderstanding of what the Overton window represents. People argue that you should suppress ideas you disagree with so that the measurement of the Overton window shows an opinion that is under-sampled against your adversary and consequently moves in the direction you prefer. This one of the most damaging examples of Goodhart's law that I know of.
To stick with the music analogy, I think if Guns 'n' Roses appeared before the Beatles there would have been a significant negative response from the public (although I would really like to pull an open minded musical expert out of history to capture their experiences of modern music). Some experts favour protecting the establishment, while others are the very first to realise the significance of a revolutionary new thing.
People are generally repelled by objectionable views and while the Overton window suggests that the notion of what is objectionable might change over time, suppressing objectionable views removes that repulsion from them while simultaneously being an act that many find objectionable. Both changes cause the dominant public opinion to move in the same direction, the opposite to what the people attempting to control the dialog desire. At the same time making the Overton window harder to measure, obscuring their failure.
The decline of deviance could be thought of as either a shrinking or expansion of the standard deviation of the Overton window. It depends upon your perspective and if you consider objective measures of variance to be more significant to subjective measures.
When the Overton window is much wider, there are a much broader set of opinions in the world, but also, by definition with the same level of acceptance as a compressed window. everything within the window is accepted. You could interpret that as a decline in deviance because you just don't consider the range of things accepted to be deviant.
When the Overton window is narrow, social pressures cause people to restrict their behaviour, which would also be considered a decline in deviance. On the other hand it would take much less to be considered deviant.
This makes me wonder if you need a second order Overton to measure the acceptability of opinions relative to their proportionate position on the Overton window. Would such a measurement measure polarisation? I would imagine that the ideal arrangement, no matter what the width of the Overton window was, would be a slower decline in acceptance of things that are disagreed with.
Once again though. If you started measuring this, would it become a target, and subject to gaming?