I would definitely believe that some parties in the advertising ecosystem are ignoring opt outs, but I'm not convinced that this methodology demonstrates that.

I would definitely believe that some parties in the advertising ecosystem are ignoring opt outs, but I'm not convinced that this methodology demonstrates that. The problem is, I'm not seeing anywhere where they describe setting up a history for these simulated users. For example, if they had said that they first navigated the user to a site about mattresses, cars, or some other highly marketable category, then I would expect to see much lower bids when opt outs were respected. But if they aren't doing anything like that then it isn't surprising that an opt-out fails to decrease bids, because there's nothing to be targeting on. Their framework is quite close, however, to something that really could do a good job here. Compare these cases: * Visit some sites that do not indicate commercial interest, then compare bidding behavior opt-in and opt out. That's what they did, and you shouldn't see much of a difference. * Visit some sites that indicate specific commercial interest, then compare bidding behavior opt-in and opt out. You should see higher bids and ads that are related to that commercial interest for the opt-in category. If you don't, something went wrong. If you see those same higher bids in the opt-out case, then consent is not being respected. (I used to work in an adjacent area; speaking only for myself)