Are you speaking from experience and/or data, or based on logical reasoning? I'm honestly curious, so I hope this doesn't come across as a digital call-out of some kind.
To elaborate, I feel like a lot of people don't pay for news, or written content, because they expect (and can get it) for free. That doesn't mean they're not good potential widget customers.
I'm curious as to what kind of inferences advertisers make / have made about users based on their specific spending patterns. For example, while I'm sure that generally speaking someone who pays for service X may be more likely than average to pay for service Y, I would also expect that variance to grow or shrink a lot depending on the specific circumstances.
> I feel like a lot of people don't pay for news ... That doesn't mean they're not good potential widget customers.
I'm having a hard time thinking of any advertiser who would be happy about excluding the sort of dedicated reader who is willing to pay for content online. I'm not saying the remaining ad impressions are "bad" just that they are objectively less valuable. That's why this scheme disproportionately devalues ad impressions: you remove the direct impressions from people who buy a pass and you also make the remaining impressions less valuable.
Look at how washingtonpost.com positions itself to advertisers: http://www.washingtonpost.com/wp-adv/mediacenter/html/ad_was... and then imagine how much harder it would be to buy into that story if it excluded many wealthy and loyal readers.
In practice, I think you'd be limited to selling network ads, which don't pay very much and rely on tracking users and building profiles to serve them targeted ads.