I’m not asking for a Facebook clone I just want to be able to have friends and make lists for them and they make lists for me.
Most such interaction now occurs within extant social groups. Much as you'd watch a TV show or go to a theatre with friends, most people now discuss films on their self-centric social media. FB / Instagram / Reddit / Twitter, etc., usually with at least some people they know directly.
Conversing with absolute strangers has far less appeal.
A decade and a half ago, its reach was greater than Netflix's.
My point wasn't the scale of the hosting network but the failure of broadcaster- (or program-originator-) centric "interactive" television concepts. I think the trope existed even when Neil Postman was writing his books ... not sure if this shows up in Amusing Ourselves to Death or Technopoly, or both, though I think it does. They're 1980s / 1990s. Postman died in 2003 though, so if it was him, it was early in Internet time.
Wikipedia lists attempts dating to the 1950s, though these were not networked interactions. Those existed, however, by 1977.
If implemented now, what about in 2032 when you're watching re-runs from 2022?
What if the store goes bust or styles change?
Display advertising will show you products that you have expressed an interest in today.
Streaming video has ad slots similar to traditional linear TV because it's what people are used to. There is the opportunity to do much much more.
What I could see happening is content changing to be tailored to the viewer. This would both decrease subscriber churn in the increasingly competitive streaming space, as well as add a feature to advertise via the content itself.
You like the pants? Point your remote at them to purchase with a single click. Honestly, I could see Amazon sponsoring streaming content for that purpose. AWS has streaming services and Amazon has been doing the logistics for decades.
I envision clothing & hairstyles being changed in the same video from one year to another, or even customized per viewer demographic. Admins could simply toss some product links over and let the ML do its thing.
Of course, this would require a ton of standardization of product types and categories, which I could see being difficult to get consensus on. Perhaps one company/consortium would corner the market and just use human labor to fill the gaps where automation isn't ready, much as is often done today.
Spam would be a huge problem, of course.