Question 1:
In working with ad serving vendors, many give great acolades to the view through convert - those that take action without having ever clicked through. My question is - if you have a attribution model that always uses last click within some time window (say 30 days), then in a perfect world - if you paid $1.00 CPM to reach 200 million online US consumers a month (or $200k) then wouldn't this methodology attribute all conversion to a served ad unless some other direct response mechanism could be determined (email, search, affiliates, etc.). So for a little over $2 million a year - you could attribute all your conversions to banner ad view through.

Question 2:
Many ad agencies have their clients run ads on remnant networks - because of the efficiency. But let's assume that you run a blinded study using a house ad (red cross, etc.) and a real advertisingment and there is some % difference in attributed response. Why then, would it not be just as reasonable to assume that if you create a ad with an ad placement on a behaviorial network or direct buy that has greater CTR than one of the remnant networks - that there would be an incremental % to those placements that could also be attributed to view through - perhaps even higher view through since the view was in context to something related to the ad in the first place.

Its essential to make sense of impression based conversions that your tracking architecture is able to dedupe effectively across all the media channels that you use .

Media often will have the effect of increasing searches for brands or offers and being able to identify the patterns between channels and cross channel impact is increasingly important